Acute upper gastrointestinal bleeding (UGIB) is a common gastrointestinal disease emergency and a cause of morbidity and mortality.
To assess the clinical outcomes and explore predictive factors for mortality of elderly patients with acute UGIB.
During the study period from January 2010 to September 2011, we prospectively enrolled 981 patients presenting with UGIB from 11 hospitals (mean age ± standard deviation (SD), 59.4 ± 14.9 years; range, 17–94 years; including 661 men). Of these 981 patients, 499 (50.9%) were elderly. Basic demographic data and clinical findings, and Rockall scores were collected and calculated.
We studied 499 elderly patients. Their mean age ± SD was 71.63 ± 7.65 years. The 30-day mortality rate was 9% and rebleeding was just 1%. Regression analysis showed a pulse rate >100 beats per min at first visit, red blood in a nasogastric aspiration, comorbidity with coronary artery disease, and creatinine >1.5 mg/dL were independent predictive factors of 30-day mortality.
Peptic ulcer bleeding is a major cause of acute UGBI in the elderly. We recommend patients with predictive factors of mortality, pulse rate >100 beats per min at first visit, red blood in nasogastric aspiration, comorbidity with coronary artery disease, and creatinine >1.5 mg/dL be closely monitored and treated promptly. Reducing mortality from peptic ulcer bleeding should focus on preventing peptic ulcer occurrence as a result of ulcerogenic medications.
Acute upper gastrointestinal bleeding (UGIB) is a common gastrointestinal emergency and cause of hospital admission. Its prevalence ranges between 37-172 per 100,000 adult persons and mortality ranges between 3% and 14% . The incidence of UGIB has decreased significantly in the past 10 years, but mortality and rebleeding rates have not , and the mortality rate has increased in elderly patients. Initial assessment and rapid resuscitation of patients with UGIB are important from the outset. Further investigations should be conducted after stabilization. Various scoring systems have been developed to identify and stratify patients from different risk levels [3-5]. The commonly used scoring systems are the Rockall and Blatchford systems, the Baylor bleeding score, and the Cedars-Sinai Medical Center Predictive Index . The challenge in caring for patients is how to identify those who are at risk of rebleeding and death. The Rockall score is the best-known risk stratification tool for patients with UGIB to predict rebleeding and mortality . The Rockall score consists of patient age, hemodynamic parameters, comorbidities, and endoscopic findings. A strength of the Rockall score lies in its simplicity. It can be measured soon after presentation with UGIB as a clinical Rockall score or can be measured completely after endoscopy as a complete Rockall score . The Rockall score has been validated in many countries and used to predict outcomes [7-10]. Thailand is confronted with an aging society. Although age has been considered as predictive of an adverse outcome of UGIB in several studies [3, 4, 11, 12], one study in Thailand did not find that advanced age influenced adverse outcomes . Every patient, but especially elderly patients, should be stratified for risk at first presentation with UGIB. Predictive factors should be evaluated to estimate mortality in elderly patients with UGIB.
The aim of the present study was to determine the Rockall score and other factors predicting outcome of elderly patients with UGIB.
Material and methods
The study was approved by the institutional review board of each hospital involved. The hospitals included King Chulalongkorn Memorial Hospital (approval No. 480/51), Maharat Nakhonratchasima Hospital (approval No. 0027.124/443), Sawanpracharak Hospital (approval No.1/2552), Surin Hospital (approval No. 0027.102/12475), Maharaj Nakhonsithammarat Hospital (approval No. 002/2555), HRH Princess Maha Chakri Sirindhorn Medical Center (approval number 49/2552 SWUEC 5-5/2552), Chonburi Hospital (approval No. 23/2553), Bangkok Metropolitan Administration General Hospital (approval No. W68), Bangkok Hospital (approval 23/12/2552), Rajavithi Hospital (approval No. 52116), and Thammasat University Hospital (approval No. MTU-E-1-78/ 52).Written informed consent was obtained from each patient or their nearest relative before enrollment at each center.
We prospectively enrolled 981 patients with UGIB from 11 hospitals in Thailand during the study period from January 2010 to September 2011. The mean age ± SD of the patients was 59.4 ± 14.9 years (range 17 - 94 years) and they included 661 men (67%). Of 981 patients, 499 (50.9%) were elderly with UGIB. Management of UGIB was based on the Consensus for Clinical Practice Guidelines for the management of UGIB from the Thai Association for Gastrointestinal Endoscopy. All patients presenting to the emergency or outpatient departments with acute UGIB were clinically assessed initially and hemodynamically stabilized. Gastroenterologists were informed. Every patient was admitted to hospital and an esophagogastroduodenoscopy (EGD) was performed according to a standardized examination protocol. If endoscopy was not available, the patient was considered for referral. Clinical management, including medication, fluid resuscitation, and the timing of endoscopy and endoscopic therapy was decided by each individual gastroenterologist. Packed erythrocyte (PRC) transfusion was considered when hemoglobin was less than 9 g/dL. If endoscopic treatment failed, the rescue therapy was surgery or embolization. The Rockall score is the best-known of all tools for risk stratification of patients with acute UGIB. It was formulated by medical institutions in the England in 1996 and its primary negative outcome is mortality . The Rockall score consists of 5 variables and is the most widely used and validated tool to predict outcome in clinical practice . In the present study, we used complete Rockall scores to predict outcome in elderly patients with acute UGIB and search for other predictive factors.
Definitions and outcomes
UGIB was defined as hematemesis (including “coffee grounds” vomiting), melena, and hematochezia. All patients underwent EGD during the study period. The definition of elderly patient was “age more than 60 years”.
The primary outcome was 30 day mortality and its predictor, which included all causes with mortality occurring within the 30 days of hospitalization.
High risk stigmata from endoscopic findings were characterized as adherent clots, nonbleeding, or bleeding from visible vessels and varices with a red color or a white nipple sign. Rebleeding was defined as a new episode of objective evidence of UGIB, after the initial bleeding had stopped, with a decreased hemoglobin concentration of at least 2 g/dL per day, or with the need for more than two units of PRC per day, or with circulatory instability.
IBM SPSS Statistics for Windows (version 19.0; IBM Corp, Armonk, NY, USA) was used for data analysis. Rockall score of all patients were calculated based on individual variables and summarized as complete Rockall score. Continuous variables were compared using a t test and categorical variables were compared using a chi square or Fisher exact test. Logistic regression analysis was conducted to determine independent predictors of 30-day mortality. A two-sided P < 0.05 was considered significant.
We included 499 elderly patients (300 male, 199 female) with acute UGIB in this study. Their age range was from 60 to 94 years. Most of the patients presented with hematemesis. Comorbidities were found in 331 patients including hypertension, diabetes mellitus, cardiovascular disease, renal disease, and cirrhosis. Antiplatelet drugs (aspirin, clopidogrel), nonsteroidal anti-inflammatory drugs (NSAIDs), and warfarin were used regularly in 171 patients (Table 1). EGD was performed within 12-48 h of admission. The vast majority of the patients (98%) were given a proton-pump inhibitor orally, intermittently or continuously intravenously, and in cases with suspected variceal bleeding, vasoactive drugs were given. Interventions were performed in high risk lesion of peptic ulcer bleeding group with a heater probe or argon plasma coagulation in 47 cases and a hemoclip in 28 cases. Cyanoacrylate glue injections were applied in 9 cases for treatment of gastric varices and esophageal varix ligations were applied in 35 cases for esophageal variceal bleeding.
Characteristics of 499 elderly patients presenting with acute upper gastrointestinal bleeding
|Variable [number (%) or mean ± SD]||Total (n = 499)|
|Age (year)||71.6 ± 7.65|
|Male: n (%)||300 (60.1%)|
|Presenting symptom: n (%)|
|Clinical findings: n (%)|
|Pulse rate >100 beats/min||26 (5.2%)|
|Systolic blood pressure <100 mmHg||26 (5.2%)|
|Presence of hemodynamic instability||71 (14.2%)|
|(pulse >100 bpm and systolic blood|
|pressure <100 mmHg)|
|Nasogastric lavage findings: n (%)|
|Red blood on nasogastric lavage||60 (12%)|
|“Coffee grounds”||181 (36.3%)|
|Normal content||144 (28.8%)|
|No nasogastric lavage||114 (22.8%)|
|Diabetes mellitus||125 (25%)|
|Cardiovascular disease||44 (8.8%)|
|Chronic renal failure||37 (7.4%)|
|Alcohol drinking: n (%)||73 (14.6%)|
|Smoking: n (%)||80 (16%)|
|Previous use of medication: n (%)|
|Low-dose aspirin||86 (17.2%)|
|Non-steroidal anti-inflammatory drugs||77 (15.4%)|
|Laboratory features at presentation|
|Hemoglobin (g/dL)||8.70 ± 3.3|
|White blood count (103/μL)||11.6 ± 12.9|
|Platelets (103/μL)||218.7 ± 115.5|
|International normalized ratio of prothrombin time||1.2 ± 0.8|
|Blood urea nitrogen (mg/dL)||32.8 ± 23.1|
|Creatinine (mg/dL)||1.7 ± 1.5|
|Pack red cell: n (%)||354 (70.9%)|
|Fresh frozen plasma: n (%)||58 (11.6%)|
|Platelets: n (%)||35 (7%)|
|Proton pump inhibitor: n (%)|
|Oral route||25 (5%)|
|Intravenous bolus||392 (78.5%)|
|Intravenous continuous drip||72 (14.4%)|
The most common cause of acute UGIB was peptic ulcer bleeding (Table 2). The 30 day mortality in elderly patients with UGIB was 9% and rebleeding after initial endoscopy was found in only 1% (Table 3). There was no significance difference in mortality between the nonelderly and elderly group (data not shown). The Rockall score was calculated based on individual data and divided into 3 risk groups (Table 4).
Endoscopic findings in 499 elderly patients presenting with acute upper gastrointestinal bleeding
|Variable||Total (n = 499)|
|Peptic ulcer as source of bleeding: n (%)|
|Clean base||263 (52.7%)|
|Pigment spot||37 (7.4%)|
|Adherent clot||8 (1.6%)|
|Nonbleeding visible vessel||51 (10.2%)|
|Bleeding visible vessel||36 (7.2%)|
|Variceal bleeding as source of bleeding: n (%)|
|Esophageal varices (EV)||41 (8.2%)|
|Gastric varices (GV)||2 (0.4%)|
|EV and GV||16 (3.2%)|
|Portal hypertensive gastropathy||46 (9.2%)|
|Other Endoscopic findings (%)|
|Mallory–Weiss tear||8 (1.6%)|
|Ulcerative mass||5 (1%)|
|No lesion identified||3 (0.6%)|
Clinical outcomes of elderly patients presenting with acute upper gastrointestinal bleeding
|Variable||N = 499|
|Endoscopic therapy for bleeding peptic ulcer: n|
|Endoscopic therapy for variceal bleeding: n|
|Endoscopic band ligation||35|
|Hospital stay: days||13.9|
|Rebleeding within 30 days: n (%)||5 (1%)|
|In hospital mortality: n (%)||17 (3.4%)|
|30-day mortality: n (%)||45 (9%)|
|Surgery or intervention: n (%)||6 (1.2%)|
Rockall score stratification into 3 groups with percentage of rebleeding and mortality of elderly patients presenting with acute upper gastrointestinal bleeding
|Group||Score||Cases (%)||Rebleeding (%)||Mortality (%)|
|Low-risk||<3||244 (48.9%)||1 (0.2%)||21 (8.6%)|
|Moderate-risk||3-4||199 (39.9%)||3 (0.60%)||18 (9.0%)|
|High-risk||>4||56 (11.2%)||1 (0.2%)||6 (11%)|
Subgroup analysis of patients with or without survival outcome
In hospital mortality and 1 month mortality were 3.4% and 9%, respectively. The common causes of death were sepsis and multiorgan failure. Mean age, hemoglobin, blood urea nitrogen, and creatinine levels were significantly higher in patients in the nonsurvival group than in the survival group (Table 5). The following variables had significant influence on predicting mortality of elderly patients by univariate analysis: comorbidities with coronary artery disease (CAD), chronic kidney disease, hemodynamic instability, pulse >100 beats per min, red blood in nasogastric lavage, fresh-frozen plasma transfusion, endoscopic finding of esophageal and gastric varices, and glue injection (Table 6). Multivariate analysis showed that mortality was significantly associated with hemodynamic instability at presentation (pulse rate >100 beats per min), comorbidity with CAD, red blood in nasogastric lavage, and creatinine more than 1.5 mg/dL (Table 7).
Predictor of 30-day mortality by univariate analysis (continuous variables) of 499 elderly patients presenting with UGIB
|Predictors||Death (n = 45)||Survival group (n = 454)||P|
|Mean age (± SD) in years||74.1 ± 7.3||71.4 ± 7. 7||0.01|
|Hemoglobin (g/dL)||9.7 ± 5.5||8.6 ± 3||0.02|
|Platelet count||213.9 ± 134.1||219.2 ± 113.6||0.38|
|International normalized ratio of prothrombin time||1.4 ± 1.3||1.2 ± 0 . 8||0.06|
|Blood urea nitrogen||40.7 ± 23.6||32 ± 22.9||0.008|
|Creatinine (mg/dL)||2.5 ± 2.1||1.6 ± 1 . 5||0.004|
|Rockall score||2.9 ± 1.7||2.7 ± 1.5||0.4|
Predictors of 30 day mortality of 499 elderly patients presenting with acute upper gastrointestinal bleeding on univariate analysis (categorical variables)
|Predictors||No. of deaths/No. of patients with predictors||No. of deaths/No. of patients without predictors||P||Odds Ratio (95% CI)|
|Coronary artery disease||8/44 (18%)||37/455 (8%)||0.048||2.49 (1.1 -5.7)|
|Chronic kidney disease||7/37 (19%)||38/462(8%)||0.03||2.58 (1.1 -6.3)|
|Presence of hemodynamic instability||12/62 (19%)||33/437 (8%)||0.003||2.9 (1.4 -5.9)|
|(pulse >100 bpm and SBP <100 mmHg)|
|Pulse rate >100 bpm||8/26 (31%)||37/473 (8%)||0.001||5.19 (2.1-12.7)|
|Red blood in NGT||18/60 (30%)||27/43 (6%)||<0.001||6.48 (3.3-12.7)|
|FFP transfusion||11/58 (19%)||34/441 (8%)||0.006||2.73 (1.3 -5.7)|
|EV and GV finding||4/16 (25%)||41/433 (10%)||0.048||3.56 (1.1-11.5)|
|Cyanoacrylate glue injection||4/9 (44%)||41/490 (8%)||0.005||8.68 (2.2-33.6)|
Independent predictors of 30 day mortality (logistic regression analysis)
|Predictors||P||Odds Ratio||95% CI|
|Pulse rate >100 beats/min||0.019||3.91||1.2–12.2|
|Red blood in nasogastric tube||<0.001||5.44||2.3–12.6|
|Creatinine more than 1.5||0.002||3.13||1.5–6.5|
|Coronary artery disease||0.028||3.27||1.1–9.4|
The main findings of the present study were factors predictive for mortality of elderly patients with acute UGIB. The elderly are a special population who are prone to have more complications from any illness because of their physiological instability, and from comorbidities and medications. In this study, we focused on elderly patients with acute UGIB, in terms of mortality and factors predictive of mortality. The cause of UGIB was peptic ulcer bleeding in 84.8% of patients, which is consistent with previous studies [13-16]. The cause of peptic ulcer bleeding might be from ulcerogenic medications (including aspirin, clopidogrel, NSAIDs, and warfarin). Rebleeding in this study was found to be around 1%; very low compared with a previous study . Mortality was found to be 9% as consistent with previous studies [12-13], but lower than in some studies that found high mortality in the elderly ranging from 12% to 25% [17-19]. An extremely elderly population had mortality rates higher than a moderately elderly group (13.6% and 8.1% respectively), with no significant difference between them (P = 0.1). The low mortality found in the present study may be explained by new devices used to stop bleeding, and the skill of the physicians as exemplified by the low rate of surgical rescue and invasive radiological interventions (1.6%) reported. The leading causes of death were sepsis and multiorgan failure.
In the present study, there were no differences in mortality among elderly patients with UGIB and low, moderate, or high Rockall scores, which differs from earlier research that found the Rockall score to be clinically useful and accurate in predicting rebleeding, rescue surgery, and mortality in elderly patients , and patients in all age groups [20-23]. However, if we look at individual factors in the Rockall score, the results are different in univariate analysis. The present study found pulse rate >100 beats per min, comorbidities with CAD or chronic kidney disease, and endoscopic findings of gastric varices were significantly associated with mortality.
Factors predictive of mortality are clinically important in the management of disease because patients with such factors should be evaluated first and repeatedly. We found that independent factors predictive of mortality were a pulse rate of >100 beats per min, red blood in the nasogastric tube, comorbidity with coronary artery disease, and creatinine more than 1.5 mg/dL. This information is important in clinical practice because these factors in patients should alert doctors to perform prompt resuscitation, urgent endoscopic diagnosis, and treatment to reduce mortality. Previous studies found early predictors of adverse prognosis in acute UGIB in the elderly, were shock, severe comorbidities, red blood emesis, variceal bleeding, gastric ulcer larger than 2 cm, and units of blood transfused [11, 24-28]. Rapid clinical triage decisions in emergency acute UGIB is crucial. The most common cause of acute UGIB was peptic ulcer bleeding (gastric or duodenal ulcer), which is not surprising considering Helicobacter pylori infection is found in 48.2% of the Thai population . Explanations for the cause of peptic ulcer bleeding are common concomitant use of aspirin, clopidogrel for prevention of atherosclerosis, and excessive use of NSAIDs for musculoskeletal pain.
The strength of this study is the use of sample data from multiple centers in four regions of Thailand, which reflect authentic life clinical situations and practices. There are some limitations in this study. First, the decision regarding the endoscopic therapy and blood transfusion is subjective, and there may be variability between gastroenterologists among the 11 centers. Second, emergency endoscopies were not available in several of the study centers. Third, patients with UGIB managed as outpatients were not enrolled in this study.
The 30 day mortality in elderly patients with UGIB was found to be 9%. The most common cause was peptic ulcer bleeding. Factors predictive of mortality were found to be a pulse rate of more than 100 beats per min on first visit, red blood cells in nasogastric aspiration, and comorbidities, such as having CAD or a creatinine level of more than 1.5 mg/dL. These factors should be taken into consideration when triaging high risk patients for immediate attention, resuscitation, close monitoring, and early treatment.
This research study was supported by the National Research Council of Thailand (NRCT), the Clinical Research Collaboration Network, Thailand (CRCN-2553-n-03), and Research Unit of Hepatitis and Liver Cancer, Chulalongkorn University, Bangkok, Thailand.
van Leerdam ME. Epidemiology of acute upper gastrointestinal bleeding. Best Pract Res Clin Gastroenterol. 2008; 22: 209-24.
van Leerdam ME Vreeburg EM Rauws EA Geraedts AA Tijssen JG Reitsma JB et al. Acute upper GI bleeding: did anything change? Time trend analysis of incidence and outcome of acute upper GI bleeding between 1993/1994 and 2000. Am J Gastroenterol. 2003; 98:1494-9.
- Export Citation
van Leerdam ME, Vreeburg EM, Rauws EA, Geraedts AA, Tijssen JG, Reitsma JB, et al. Acute upper GI bleeding: did anything change? Time trend analysis of incidence and outcome of acute upper GI bleeding between 1993/1994 and 2000. Am J Gastroenterol. 2003; 98:1494-9.)| false 10.1111/j.1572-0241.2003.07517.x
Rockall TA Logan RFA Devlin HB Northfield TC. and the steering committee and members of the National Audit of Acute Upper Gastrointestinal Haemorrhage. Risk assessment after acute upper gastrointestinal haemorrhage. Gut. 1996; 38:316-21.
Blatchford O Murray WR Blatchford M. A risk score to predict need for treatment for upper gastrointestinal haemorrhage. Lancet. 2000; 356:1318-21.
Das A Wong RC. Prediction of outcome of acute GI hemorrhage: a review of risk scores and predictive models. Gastrointest Endosc. 2004; 60:85-93.
Stanley AJ. Update on risk scoring systems for patients with upper gastrointestinal haemorrhage. World J Gastroenterol. 2012; 18:2739-44.
Camellini L Merighi A Pagnini C Azzolini F Guazzetti S Scarcelli A et al. Comparison of three different risk scoring systems in non-variceal upper gastrointestinal bleeding. Dig Liver Dis. 2004; 36:271-7.
Vreeburg EM Terwee CB Snel P Rauws EA Bartelsman JF Meulen JH et al. Validation of the Rockall risk scoring system in upper gastrointestinal bleeding. Gut. 1999; 44:331-5.
Enns RA Gagnon YM Barkun AN Armstrong D Gregor JC Fedorak RN. Validation of the Rockall scoring system for outcomes from non-variceal upper gastrointestinal bleeding in Canadian setting. World J Gastroenterol. 2006; 12:7779-85.
Church NI Dallal HJ Masson J Mowat NA Johnson DA Radin E et al. Validity of the Rockall scoring system after endoscopic therapy for bleeding peptic ulcer: a prospective cohort study. Gastrointest Endosc. 2006; 63:606-12.
Marmo R Koch M Cipolletta L Capurso L Pera A Bianco MA et al. Predictive factors of mortality from non-variceal upper gastrointestinal hemorrhage: a multicenter study. Am J Gastroenterol. 2008; 103: 1639-47.
Wang CY Qin J Wang J Sun CY Cao T Zhu DD. Rockall score in predicting outcomes of elderly patients with acute upper gastrointestinal bleeding. World J Gastroenterol. 2013; 19:3466-72.
Charatcharoenwittaya P Pausawasdi N Laosanguaneak N Bubthamala J Tanwandee T Leelakusolvong S. Characteristics and outcomes of acute upper gastrointestinal bleeding after therapeutic endoscopy in elderly. World J Gastroenterol. 2011; 17: 3724-32.
- Export Citation
Charatcharoenwittaya P, Pausawasdi N, Laosanguaneak N, Bubthamala J, Tanwandee T, Leelakusolvong S. Characteristics and outcomes of acute upper gastrointestinal bleeding after therapeutic endoscopy in elderly. World J Gastroenterol. 2011; 17: 3724-32.)| false 10.3748/wjg.v17.i32.3724
Cooper BT Weston CF Neumann CS. Acute upper gastrointestinal hemorrhage in patients aged 80 years or more. Q J Med. 1988; 68:765-74.
Nakhonya JM Datta-Chaudhuri ML Bhan GL. Acute upper gastrointestinal hemorrhage in older people: a prospective study in two neighboring districts. J Am Geriatr Soc. 1997; 45:752-4.
Segal WN Cello JP. Hemorrhage in the upper gastrointestinal tract in the older patient. Am J Gastroenterol. 1997; 92:42-6.
Kaplan RC Heckbert SR Koepsell TD Furberg CD Polak JF Schoen RE et al. Risk factors for hospitalized gastrointestinal bleeding among older persons. Cardiovascular Health Study Investigation. J Am Geriatr Soc. 2001; 49:126-33.
Peter DJ Dougherty JM. Evaluation of the patient with gastrointestinal bleeding: an evidence based approach. Emerg Med Clin North Am. 1999; 17:239-61.
Chow LW Gertsch P Poon RT Branicki FJ. Risk factor for rebleeding and death from peptic ulcer in the very elderly. Br J Surg. 1998; 85:121-4.
Rockall TA Logan RF Devlin HB Northfield TC. Selection of patients for early discharge or outpatient care after acute upper gastrointestinal haemorrhage. National Audit of Acute Upper Gastrointestinal Haemorrhage. Lancet. 1996; 374:1138-40.
Rockall TA Logan RF Devlin HB Northfield TC. Variation in outcome after acute upper gastrointestinal haemorrhage. The National Audit of Acute Upper Gastrointestinal Haemorrhage. Lancet. 1995; 346: 346-50.
Rockall TA Logan RF Devlin HB Northfield TC. Influencing the Practice and outcome in acute upper gastrointestinal haemorrhage. Steering Committee of the National Audit of Acute Upper Gastrointestinal Haemorrhage. Gut. 1997; 41:606-11.
Lahiff C Shield W Cretu I Mahmud N McKiernan S Norris S et al. Upper gastrointestinal bleeding: predictors of risk in a mixed patient group including variceal and nonvariceal haemorrhage. Eur J Gastroenterol Hepatol. 2012; 24:149-54.
Mondardini A Barletti C Rocca G Garripoli A Sanbataro A Perotto C et al. Non-variceal upper gastrointestinal bleeding and Forrest’s classification: diagnosis agreement between endoscopists from the same area. Endoscopy. 1998; 30:508-12.
Himal HS Watson WW Jones CW Miller L Maclean LD. The management of upper gastrointestinal hemorrhage: a multiparametric computer analysis. Ann Surg. 1974; 179:489-93.
Sung JJ Tsoi KK Ma TK Yung MY Lau JY Chiu PW. Cause of mortality in patients with peptic ulcer bleeding; a prospective cohort study of 10428 cases. Am J Gastroenterol. 2010; 105:84-9.
Chiu PW Ng EK Cheung FK Chan FK Leung WK Wu JC et al. Predicting mortality in patients with bleeding peptic ulcers after therapeutic endoscopy. Clin Gastroenterol Hepatol. 2009; 7:311-6.
Zimmerman J Siguencia J Tsvang E Beeri R Arnon R. Predictors of mortality in patients admitted to hospital for acute upper gastrointestinal hemorrhage. Scand J Gastroenterol. 1995; 30:327-31.
Atisook K Kachinthorn U Luengrojanakul P Tanwandee T Pakdirat P Puapairoj A. Histology of gastritis and Helicobacter pylori infection in Thailand: a nationwide study of 3776 cases. Helicobacter. 2003; 8:132-41.