Epidemiology and Treatment of Colonic Angiodysplasia; a Population-Based Study. Naomi G. Diggs, MD Lisa L. Strate, MD MPH March 2, 2010
Background Angiodysplasia is an important cause of occult and acute GI bleeding, especially in the elderly. Pathophysiology Obstruction of submucosal veins Decreased mucosal blood flow and local ischemia Prevalence estimates vary widely Asymptomatic people <1% Patients with hematochezia 37-50% Boley et al, Gastro 1977 Foutch et al, Am J Gastro 1995 Jensen et al, Gastro 1988 Santos et al, Br J Surg 1988
Background Clinical Features Older patients (>60 years) Frequently R-sided colonic lesions but present throughout GI tract Occult or acute blood loss Endoscopic diagnosis and therapy Colonoscopy is 80% sensitive Therapy is widely accepted and successful Indicated for evidence of blood loss or actively bleeding lesions Richter et al, Dig Dis Sci 1984 ASGE Guidelines, Gastro Endosc 2005
Study Aims To use the Clinical Outcomes Research Initiative (CORI) database to study: 1. Epidemiology of colonic angiodysplasia 2. Risk factors for occult or overt bleeding 3. Nationwide trends in treatment practices
Clinical Outcomes Research Initiative (CORI) Data repository at OHSU developed to study endoscopic procedures. Consortium of gastroenterology practices 73 sites from 27 states Community 71% University 17% VA 12% 525 participating physicians Endoscopic report generator Patient demographics, lesions found, treatment, etc
Study Population 229,727 colonoscopies performed between Jan 2000 to Dec 2002. All unique patients with at least one angiodysplasia (n=4,159) Exclusion Criteria: Age <20 years, radiation proctitis, angiodysplasia of rectum only (n=103). Data collected Patient demographics, ASA class, bleeding indications Exam location, practice type Angiodysplasia number, size, location, bleeding Treatment
Statistical Analysis Outcomes of interest 1. Evidence of overt or occult blood loss - Fe deficiency anemia - Positive FOBT - Hematochezia - Melena - Active bleeding at time of procedure 2. Receipt of endoscopic treatment Multivariate logistic regression to identify predictors of each outcome.
Patient Demographics Total Patients n= 4,159 Age (years) 20-39 2.5% 40-59 25% 60-79 58% >80 15% Male 55% Caucasian 86% ASA class > III 16% Outpatient 84% Community -based practice 69%
Exam Indications Total Patients n= 4,159 Evidence of Blood Loss 2,320 (56%) Hematochezia 41% Anemia 37% Positive FOBT 33% Melena 7% No bleeding 1,839 (46%) Other Findings Diverticula 54% Polyps 40% Colitis 4% Malignancy 0.8%
Angiodysplasia Characterisitics Characteristic n= 4,616 Lesion number Unique (1) 41% Multiple (2-10) 30% Diffuse (>10) 9% Size (mm) Minute (<2) 3% Intermediate (2-5) 70% Large (>5) 19% Active Bleeding 7%
Location of Lesions Transverse Colon 7% Right Colon 62% Left Colon 30%
Predictors of Blood Loss Characteristic OR (95%CI) p value Inpatient exam 9.6 (6.0-15.3) <0.0001 Age >80years 1.4 (1.1-1.7) 0.003 Black 2.0 (1.5-2.6) <0.0001 Hispanic 1.7 (1.3-2.2) <0.0001 ASA class > III 1.9 (1.6-2.3) <0.0001 2-10 lesions 1.5 (1.3-1.7) <0.0001 >10 lesions 2.3 (1.8-2.9) <0.0001 Size >5mm 1.3 (1.1-1.5) 0.006
Endoscopic Treatment 18.5 % of all angiodysplasias treated 28% of patients with bleeding indications 68% of actively bleeding lesions 48% of patients with hematochezia 5% of patients without bleeding indications Modalities Thermal coagulation 60% Epinephrine and/or saline injection 24% APC 20% Combination 2.5%
Predictors of Treatment Characteristic OR (95%CI) p value Active bleeding 10 (7.7-13.2) <0.0001 Evidence of blood loss 5.1 (4.1-6.4) <0.0001 Inpatient exam 2.6 (2.0-3.4) <0.0001 Age 20-39 yrs 3.1 (2.0-5.0) <0.0001 Age >80 yrs 1.6 (1.2-2.1) 0.0008 Size >5mm 1.8 (1.5-2.3) <0.0001 University 2.6 (2.1-3.2) <0.0001
Strengths: Strengths and Limitations Largest cohort of patients with angiodysplasia examined to date. All angiodysplasia not just patients with hematochezia allows identification of risk factors for bleeding. Variety of practice types and locations allows a nationwide perspective on current treatment practices. Limitations: Retrospective data limited to endoscopy encounter Misclassification bias Unable to identify prevalence of angiodysplasia in specific populations i.e. hematochezia.
Conclusions Colonic Angiodysplasia are uncommon. Seen primarily in elderly patients. Majority of lesions are Right-sided. More than half of patients with lesions have evidence of blood loss. Large and multiple lesions more likely to be associated with blood loss. Race and ethnicity may be involved in pathogenesis or a marker of comorbidity. Endoscopic treatment is inconsistent and underutilized in patients with evidence of blood loss. Standardization of treatment is needed.
Acknowledgements Lisa Strate, MD, MPH Christina Surawicz, MD Jennifer Holub, MPH and Glenn Eisen, MD at CORI