Management and Outcome Predictors in Acute Surgical Admissions for Lower Gastrointestinal Bleeding

J. Newman1, J.E.F. Fitzgerald1, S. Gupta1, A.C. von Roon1,2, H.H. Sigurdsson1, T.G. Allen-Mersh1,2
1

Dept of General Surgery, Chelsea & Westminster NHS Hospital Trust, 369 Fulham Road, London 2 Division of Surgery, Oncology, Reproductive Biology and Anaesthetics, Imperial College London

Aims
Lower gastrointestinal bleeding (LGIB) is a common emergency department presentation. Unlike upper gastrointestinal bleeding (UGIB), there are currently no commonly accepted triaging models or scoring systems to classify these patients. We assessed the utility of current triage tools (the BLEED criteria) and identified factors which could be implemented for earlier risk stratification of patients.

Conclusions
Our study has determined an independent set of prognostic factors which could be incorporated into initial triaging of patients presenting to emergency departments with LGIB. This may facilitate early identification of patients requiring more aggressive resuscitation, admission to a monitored bed and consideration for early radiological or surgical intervention, as appropriate. It may also help prevent unnecessary admissions by identifying those who can be safely investigated as an outpatient. The existing BLEED criteria were not shown to have any predictive value in this patient cohort.

Methods
Patients were identified from a prospectively maintained surgical admissions database at a central London teaching hospital. Data collected on 26 clinical factors available on initial presentation. Primary outcome criteria were severe bleeding, and adverse outcome. Severe bleeding was defined as: continued bleeding within first 24-hours, requirement of blood transfusion, decrease in haematocrit 20% or recurrent bleeding after 24-hours of stability. Adverse outcome was defined as: emergency surgery to control bleeding, ITU admission, or death

Table 1. Independent predictors for severe LGIB

Parameter Hct <0.35 Bright red blood PR Age >60 years

OR 4.7 3.5 2.3

95% CI 1.7 13.0 1.7 7.1 1.05 - 4.9

P Value .002 <0.001 .03

Results
184 patients were admitted with LGIB during the study period (3% of all surgical referrals) Twelve patients with UGIB were excluded. Severe bleeding occurred in 110 patients (64%). Adverse outcome was recorded in 20 patients (11.6%); 10 patients (5.4%) died during admission. The commonest aetiologies were diverticular disease, haemorrhoids and malignancy. On admission three independent prognostic factors for severe bleeding were identified (Table 1): haematocrit < 0.35 (p=0.002), bright red blood per rectum on examination (p<0.001), and age >60 years (p=0.03); area under ROC 0.66 (95% CI 0.58 0.74). Four independent prognosticators of an adverse outcome were identified (Table 2): creatinine >150 mol/l (p=0.002), age >60 years (p=0.001), abnormal haemodynamic parameters on presentation (p=0.05) and continued bleeding within the first 24-hours (p=0.05); area under ROC 0.79 (95% CI 0.70 0.90). Anticoagulation use was not correlated with severe bleed or adverse outcome. The BLEED criterion was not shown to have any predictive value in this patient cohort.

Table 2. Independent risk factors for adverse outcome from LGIB

Parameter Creatinine >150 mol/l Age >60 years Abnormal haemodynamic parameters Re-bleed Smoking

OR 10.3 4.2 2.1 1.9 0.5

95% CI 2.4 43.5 1.8 10.0 1.0 4.6 1.0 3.8 0.2 1.0

P Value .002 .001 .05 .05 .05

Source of bleeding Diverticular disease Haemorrhoids Neoplasia Bleeding post polypectomy Infective colitis Ischaemic colitis Unidentified Other Upper gastrointestinal source

Total number of patients (%) 56 (32) 25 (14.5) 14 (8) 9 (5) 9 (5) 4 (2) 19 (11) 36 (21) 12

Number of patients with severe bleeding (%) 36 (64) 11 (44) 8 (57) 2 (22) 4 (44) 4 (100) 10 (53) 36 (100) Excluded

Number of patients with adverse outcome (%) 5 (9) 0 (0) 1 (7) 2 (22) 3 (33) 4 (100) 3 (16) 2 (5.5) Excluded

A comparison of receiver operating characteristics curves for the BLEED criteria , the risk stratification of patients based on data available at admission, and the risk stratification of patients based on data at 24 hours.

References
Strate LL, Orav EJ, Syngal S. Early Predictors of Severity in Acute Lower Intestinal Tract Bleeding. Arch Intern Med 2003; 163: 838-843. Das A, Ben-Menachem T, Cooper GS, Chak A, Sivak MV, Gonet JA, et al. Prediction of outcome in acute lower-gastrointestinal haemorrhage based on an artificial neural network: internal and external validation of a predictive model. Lancet 2003; 362: 1261-1266. Das A, Wong RCK. Prediction of outcome of acute GI hemorrhage: a review of risk scores and predictive models. Gastrointest Endosc 2004; 60: 85-93. Velayos FS, Williamson A, Sousa KH, Lung E, Bostrom A, Weber EJ, et al. Early predictors of severe lower gastrointestinal bleeding and adverse outcomes: A prospective study. Clin Gastroenterol Hepatol 2004; 2: 485-490. Kollef MH, O'Brien JD, Zuckerman GR, Shannon W. BLEED: A classification tool to predict outcomes in patients with acute upper and lower gastrointestinal hemorrhage. Crit Care Med 1997; 25: 1125-1132. Rockall TA, Logan RFA, Devlin HB, Northfield TC. Risk assessment after acute upper gastrointestinal haemorrhage. Gut 1996; 38: 316-321.