Michael Schopis, MD1, Arcelia Guerson Gil, MD2, Yonatan Ziv, MD1, Rie Seu, MS3, Nai Chien Yeat, MD1, Samuel Muench, MD1, Mimi Kim, PhD3, Olga Aroniadis, MD, MSc1, Lawrence Brandt, MD, MACG1
1Montefiore Medical Center, Bronx, NY; 2Jacobi Medical Center, Bronx, NY; 3Albert Einstein College of Medicine, Bronx, NY
Introduction: GI investigation of pts with A-IDA typically begins with both upper (U) and lower (L) GI endoscopy. Guidelines tend to recommend bidirectional endoscopy (BDE) sequencing based on age stratification, or starting with LGI endoscopy. The aim of this study was to determine the most efficient BDE sequence among patients with A-IDA by investigating the relative frequencies of identified associated sources of bleeding by UGI or LGI endoscopy.
Methods: We performed a retrospective chart review of clinical characteristics of pts ≥18 yrs who had BDE performed within 6 months of A-IDA diagnosis between the yrs 2014-2019. Pts were excluded if they had GI tract surgery within 10 yrs, active pregnancy, active GI bleeding, an obvious site of non-GI blood loss or ferritin >100. McNemar’s test was used to compare frequency of lesions identified across the general population and stratified for age < 50 and
>50 yrs. Bivariate analyses of characteristics associated with UGI and LGI lesions were performed using t-test, Chi-squared test, or Fisher’s exact test. Multivariate logistic regression estimated odds ratios.
Results: 497 pts were available for analysis (65% female, mean age 63 yrs; Table 1). BDE found IDA-associated lesions in 307 (61.8%) pts comprising 196 (39.4%) exclusively UGI lesions, 66 (13.3%) exclusively LGI lesions, and 45 (9.1%) both UGI and LGI lesions. In our cohort, UGI lesions were more common than LGI lesions [241 (48.5%) vs 111 (22.3%); p< 0.001], regardless of age (p< 0.001) (Table 1). Protective variables against an LGI lesion included female sex [OR 0.51 (0.32-0.86); p=0.01], BMI [OR 0.96 (0.93-0.99) per unit inc.; p=0.01] and Hgb [OR 0.88 (0.78-0.98) per unit inc.; p=0.03]. CKD was associated with LGI lesions [OR 1.99 (1.12-3.48); p=0.02]. Hgb was higher in pts with only UGI than only LGI lesions (9.4±1.6 vs 8.5±1.5; p=0.001). PPI use predicted absence of identifiable lesions [OR 0.66 (0.45-0.96); p=0.03]. The most common UGI lesions were H. pylori-associated (HP-A) gastritis (24.2%) and peptic ulcer disease (17.6%) (Fig 1a). The most common LGI lesions were large polyps (37.7%) and colon CA (16.2%) (Fig 1b).
Discussion: Our results indicate 1) pts with A-IDA are more likely to have an UGI than LGI source; thus starting with UGI endoscopy may be prudent; 2) women are less likely to have a LGI source of A-IDA than men; 3) low BMI, low Hgb and CKD predict a LGI source of A-IDA. 4) HP-A UGI lesions account for a significant portion of A-IDA; 5) PPI use predicts the absence of lesions.
Citation: Michael Schopis, MD; Arcelia Guerson Gil, MD; Yonatan Ziv, MD; Rie Seu, MS; Nai Chien Yeat, MD; Samuel Muench, MD; Mimi Kim, PhD; Olga Aroniadis, MD, MSc; Lawrence Brandt, MD, MACG. P0411 - BIDIRECTIONAL ENDOSCOPY IN ASYMPTOMATIC-IRON DEFICIENCY ANEMIA (A-IDA). Program No. P0411. ACG 2019 Annual Scientific Meeting Abstracts. San Antonio, Texas: American College of Gastroenterology.