| Abstract|| |
This retrospective analysis studied the records of 564 consecutive patients admitted to Gastrointestinal Bleeding Unit of Riyadh Medical Complex with acute upper gastrointestinal bleeding over a 2-year period (May 1996-April 1998). The purpose of the study was to analyze the mortality with an aim to identify the risk factors affecting mortality in these patients. Majority of patients were men (82%) and Saudis (54%). Their mean age was 52.46 + 17.8 years. Esophageal varices (45%) were the main causes of bleeding followed by duodenal ulcers (24%). Overall mortality in this series was 15.8% (89 patients). Comorbid diseases were responsible for death in 68 (76%) patients, whereas, bleeding was considered to be directly responsible for death in 21 (24%) patients. On analysis of data from this study, old age (>60 years), systolic pressure <90 mm Hg on admission, comorbid disease, variceal bleeding and Child's grade C in patients with chronic liver disease were associated with adverse outcome.
Keywords: Gastrointestinal bleeding, prognostic factors
|How to cite this article:|
Alam MK. Factors affecting hospital mortality in acute upper gastrointestinal bleeding. Saudi J Gastroenterol 2000;6:87-91
Acute upper gastrointestinal bleeding (AUGIB) is a common emergency. This accounts for more than 300, 000 annual hospital admissions in the United States. The management of this problem is demanding and needs a team approach. Reported mortality range from 5-15%. Despite improvements in diagnostic accuracy, the overall fatality rate from upper gastrointestinal bleeding (UGIB) has changed little in the past decades. Patients with severe medical comorbidities and those with persistent or recurrent bleeding are at highest risk of dying. Advanced age, state of shock at admission and presence of advanced chronic liver disease are other factors of adverse prognostic value,,, It has been shown that managing AUGIB in a specialized unit may improve survival.
Patients presenting with AUIGB at Riyadh Medical Complex (RMC) are admitted to a specialized gastrointestinal bleeding unit (GIBU). This is a 6-bed unit managed by general surgeons in cooperation with medical gastroenterologist. The purpose of the study was to analyze and identify factors influencing the mortality of AUGIB patients admitted to this unit. To the best of my knowledge no published local series have evaluated the impact of such factors,
| Patients and Methods|| |
The hospital records of consecutive patients with AUGIB admitted to GIBU of RMC during the 2-year period from May 1996 to April 1998 (1417H-1418) were reviewed retrospectively. Patients included in the study were those admitted to the hospital with clinical evidence of AUGIB (hematemesis or melena) or AUGIB occurring in an inpatient who had already been admitted to the hospital for different reasons. Patients in whom the bleeding was from sources other than upper gastrointestinal tract were excluded. All patients admitted to GIBU were resuscitated with crystalloid and or blood as indicated, had nasogastric lavage with ice cool saline and blood samples collected for hematological, biochemical, serologic and blood group and cross match. Upper GI endoscopy was done within 24 hours in most and within 48 hours in all who were stable to undergo this procedure. Endoscopic therapy such as sclerotherapy for bleeding varices and adrenaline injections or heat probe coagulation for bleeding ulcer was used to control active bleeding. Surgical intervention was usually indicated when endoscopic attempt failed to control bleeding or in those who rebled after initial control. Patients' demographic data, mode of presentation, endoscopic findings, comorbid disease and outcome from admission to discharge or death were recorded. Chronic liver disease was diagnosed when two of the following criterias were present: a) history of hepatitis/bilharziasis; b) clinical signs (jaundice, ascitis) of liver disease; c) deranged biochemical liver profile; and d) esophageal varices.
Histologic evidence whenever available was used as confirmation of this diagnosis. The data were analyzed to study the impact of various factors on mortality. The Chi-square test was used for statistical analysis.
| Results|| |
During the 2-year study period, 564 patients were admitted to GIBU with hematemesis and or melena. The majority were men (n=544, 82%) and Saudis (54%). The youngest patient was 12 years old;the oldest was 108 years with a mean age of 52.46 + 17.8 years. Esophageal varices (n=254, 45%), which included 82 patients with associated fundal varices, was the leading cause of bleeding. This was followed by duodenal ulcer (n=136,24%). Overall mortality in this series was 15.8% (n=89). [Table - 1] presents the relative frequency of different bleeding sources and their mortality rates. Highest mortality was observed in patients with bleeding malignant lesions. Next highest mortality was among those who could not be endoscoped because of their poor general condition. Majority from the latter group died within 48 hours from admission. Mortality among bleeding esophageal varices group was 19%. On further analysis of mortality, pre-existing medical illness was directly responsible for death in 68 patients (76%) [Table - 2]. Upper gastrointestinal bleeding in these patients was minor. However, bleeding was considered to be directly responsible for death in twenty-one (24%) patients. In two patients bleeding was excessive before reaching hospital whereas; fourteen re-bled massively after initial control and developed multi-organ failure before surgery could be performed. The remaining five patients died of complications following surgery: 3 were operated for bleeding esophageal varices, one for bleeding duodenal ulcer and one for small bowel ischaemia which developed few days after the control of bleeding oesophageal varices by sclerotherapy [Table - 2].
When other risk factors were considered, a clear relationship between increasing age and mortality was observed [Table - 3]. The majority of patients admitted with AUGIB were under 60 years of age with a mortality rate of 10%. However this rose to 29% in patients over 80 years of age group. This difference was statistically significant (P<0.0001). When the effect of hypotension on death rate was analyzed, it was found that this factor becomes significant when systolic level at presentation to hospital was at or below 90 mmHg (P <0.0001,
[Table - 4]. The effect of Child-Pugh's grading on mortality was also analyzed in patients of chronic liver disease. Complete information for this analysis was available on 196 patients. [Table - 5] presents the clear adverse effect of severe liver disease on mortality. A definite etiology of liver disease could not be established in these patients due to incomplete information.
Rebleeding occurred in 47 patients (8%). Esophageal varices were the source of rebleeding in 37 patients, a third of them bled from associated fundal varices. Duodenal ulcers were the source of rebleeding in 10 patients. Sixteen patients died of rebleeding including the two who were operated on to arrest bleeding. The majority of these patients were of older age group (average age 61 years) with underlying liver disease. Hemoglobin level on admission and number of blood unit transfused had no influence on outcome. Hemoglobin was below 9 gm/dl on admission in 168 patients (30%), 30 of them died. The remaining two-third of the patients (n=59), who died, had admission hemoglobin above 9 gm/dl. Fatality rates among patients who received zero to 4 units of blood transfusion compared to those who received 5 or more units of blood were 16.5% (n=70) and 13.5% (n=19) respectively.
| Discussion|| |
AUGIB is associated with high morbidity and mortality. If patients at higher risk of adverse outcome can be identified at an earlier stage of management the therapeutic efforts can be concentrated towards them. An attempt has been made in this study to identify those factors, which were associated with increased mortality in this population. The overall mortality in the present series was 15.8%. One previous study reported mortality close to this rate, whereas the commonly reported mortality is much lower,,. Factors which were identified with higher mortality in this series were esophageal varices as the bleeding source, poor state of liver reserve (Child's C), presence of comorbid diseases, age over 60 years, and a systolic blood pressure <90 mmHg on admission.
Bleeding esophageal varices are the commonest cause of AUGIB due to a high prevalence of chronic liver disease in this area,. High mortality in this group of patient is well documented,, Complications such as hepatic failure and rebleeding are higher in patients with poor liver reserve. Bleeding frequently aggravated liver failure while the converse is also true. All this leads to a higher mortality in these patients. Mortality rate in the present series is higher when compared to other local series. This difference can be explained by a higher proportion of bleeding esophageal varices (45% vs. 30%) and patients with advance (Child's C) liver disease (31.5% vs. 22.2%). Another risk factor which clearly adversely affected the outcome in this series were chronic medical illnesses. In 68 patients (76%), death was directly due to these long standing chronic diseases which further deteriorated in the presence of bleeding. Bleeding was minor in the majority of these patients as evidenced by only a moderate drop in hemoglobin. The adverse effect of comorbid diseases has been confirmed by several studies,. AUGIB in the elderly carries a high mortality. This is due to poor tolerance to blood loss because of limited physiological reserve and frequent significant co-existing pathology. Rebleeding can further worsen the prognosis in the elderly as seen in this study. In this series, 60% (n=53) of patients who died were over the age of sixty. The proportion of elderly patients presenting with AUGIB is generally increasing,. This trend can also be observed in this country. Al-Mofarrah et al in their study conducted more than 10 years ago reported that 21% of the patients presenting with AUGIB were 60 years of age and above. In the present study 38.5% were over the age of sixty.
Higher incidence of AUGIB in the elderly has been blamed on the increased use of nonsteroidal anti inflammatory drugs. This may be true for our population as well but there were not enough information in patients records to make a definite comment on this issue. Hypotension (systolic pressure <90 mmHg) on admission was associated with higher death rate among our patients. Holman et al found a similar level of systolic pressure to be associated with higher mortality. Other studies have also supported this finding,, In this study hemoglobin level on admission and the amount of blood transfusion did not seem to affect the outcome.
Management of AUGIB in a special unit is a logical approach where these patients are cared for by nursing staff well acquainted with the problem and medical staff expert in the field. GIBU at RMC has been established for about 15 years providing specialized care, the only one of its kind in the country. However, in the present series the benefit of such a unit may not be apparent. Mortality could have been higher with such high risk group of patients if managed in anon-specialized area.
In conclusion, on analysis of the data from the present study, old age (>60), hypotension (systolic <90 mmHg) on admission, presence of comorbid disease, variceal bleeding and child's class C liver disease are associated adverse outcome following AUGIB.
| Acknowledgement|| |
I would like to thank all the consultant surgeons of Riyadh Medical Complex for permission to study their patients. I am also thankful to Ms. Cora Rivera for her assistance in typing this manuscript.
| References|| |
|1.||Kankaria AG, Fleischer DE. The critical care management of non-variceal upper gastrointestinal bleeding. Critical Care Clinics 1995;11:347-68. [PUBMED] |
|2.||Terdiman JP. Update on upper gastrointestinal bleeding. Basing treatment decisions on patients risk level. Postgrad Med 1998;103:43-7, 51-2, 58-9. |
|3.||Katschinski B, Logan R, Davies J, Faulkner G, Pearson J, Langman M. Prognostic factors in upper gastrointestinal bleeding. Digestive Diseases and Sciences 1994,39:706-12. [PUBMED] |
|4.||Avery Jones F. Haematemesis and melena: With special reference to causation and to the factors influencing the mortality from bleeding peptic ulcer. Gastroenterology 1965:30:166-190. |
|5.||Johnston SJ, Jones PF, Kyle J, Needham CD. Epidemiology and course of gastrointestinal haemorrhage in North-East Scotland. Br M J 1973;3:655-60. |
|6.||Corley DA, Stefan AM, Wolf M, Cook EF, Lee TH. Early indicators of prognosis in upper gastrointestinal haemorrhage. Am J Gastroenterol 1998;43:336-340. |
|7.||Holman RAE, Davis M, Gough KR, Gartell P, Britton DC, Smith RB. Value of a centralized approach in the management of haematemesis and melaena: Experience in a District General Hospital. Gut 1990;31:504-8. |
|8.||Al-Mofarrah M, Fakunle YM, Al-Moagel M. Upper gastrointestinal bleeding among Saudis: Etiology and prevalence. The Riyadh Central Hospital experience. Annals of Saudi Medicine 1991;11:547-50. |
|9.||Al-Rashed RS, Laajam MA, AI-Mofleh IA, Al-Aska AK, Al-Faleh FZ. Hussein J. Acute upper gastrointestinal bleeding in King Khalid University Hospital. Annals of Saudi Medicine 1990;10:110A. |
|10.||Madden MV, Griffith GI-1. Management of upper gastrointestinal bleeding in a District General Hospital. J R Coll Physicians London 1986;20:212-15. |
|11.||Ahmed ME, Al-Knaway B, Al-Wabel AH, Malik GM, Foli AK. Acute upper gastrointestinal bleeding in Southern Saudi Arabia. J R Coll Physician London 1997:31:62-64. |
|12.||Laajam MA, Al-Mofleh IA, Al-Faleh FZ, Al-Aska AK, Jessen K, Hussain J, Al-Rashed R. Upper gastrointestinal endoscopy in Saudi Arabia: Analysis of 6386 procedures. Q J Med 1988;249:21-25. |
|13.||Jacob S, Chang RWS, Lee B et al. Prediction of outcome in patients with acute variceal haemorrhage. Br J Surg 1989;76:123-6. |
|14.||Graham DY, Smith JL. The course of patients after variceal haemorrhage. Gastroenterology 1981;80:800-9. [PUBMED] |
|15.||Merkel C, Bolognesi M. Angeli P, et al. Prognostic indicators of survival in patients with cirrhosis and oesophageal varices without previous bleeding. Ain J Gastroenterol 1989:84:717-22. |
|16.||Johnston GW, Spencer EF, Mullan FJ. Are Child's C patients with acute variceal bleeding worth treating? HPB Surg 1991;4:271-74. [PUBMED] |
|17.||Terblance J. Has sclerotherapy altered the management of patients with variceal bleeding? Am J Surg 1990:160:37-41. |
|18.||Allan R, Dykes P. A study of the factors influencing mortality rates from gastrointestinal haemorrhage. Q J Med 1976;45:533-50. [PUBMED] [FULLTEXT]|
|19.||Somerville KW, Faulkner G, Langman MJS. Nonsteroidal anti-inflammatory drugs and bleeding peptic ulcer. Lancet 1986;i:462-4. |
|20.||Lule GN, Obiero ET, Ogutu EO. Factors that influence the short term outcome of upper gastrointestinal bleeding at Kenyatta National Hospital. East Af Med J 1994;71:240-44. |
Mohammed Khurshid Alam
Department of Surgery, King Khalid University Hospital, P.O. Box 7805, Riyadh 11472
Source of Support: None, Conflict of Interest: None
[Table - 1], [Table - 2], [Table - 3], [Table - 4], [Table - 5]