PREDICTORS OF GASTRIC INTESTINAL METAPLASIA SURVEILLANCE COMPLIANCE IN A HIGH-RISK AMERICAN POPULATION

INTRODUCTION:
Countries with a high prevalence of gastric cancers have well developed programs aimed at identifying and monitoring gastric cancer precursors. However, there is limited information about the performance of such programs in the United States. In 2015, we initiated a surveillance program to follow patients with gastric intestinal metaplasia (GIM) by upper endoscopy every three years. We aim to gauge the implementation and compliance of surveillance at 7-10 years of follow-up.

METHODS:
Between March 2010 to March 2014, 366 patients had GIM on index upper endoscopy at our large metropolitan safety-net hospital. These patients were followed as part of a prospective cohort study. In our gastroenterology and primary care clinics we implemented a practice to recommend surveillance endoscopy every three years beginning in January 2015, given demographic risk factors of our population. The primary outcome was the proportion of GIM patients who returned for follow-up endoscopy with gastric biopsies. Additional outcomes were predictors of compliance, progression and regression of lesions, and length of time between surveillance.

RESULTS:
Patients identified with GIM on index upper endoscopy were predominantly Hispanic (77%) and female (62.8%), with a mean age of 57.8 years. Of this group, 90 (24.6%) patients underwent surveillance endoscopy. The average time interval was 31.8 months following the index assessment and the number of follow-up endoscopies per patient ranged from 1 to 8.

Of the 34 patients with focal intestinal metaplasia on index endoscopy, 4 (12%) progressed to multifocal metaplasia while the remainder either maintained focal metaplasia or regressed to lower risk lesions. Of the 55 patients with non-focal metaplasia initially, 5 (9%) patients progressed to neoplasia (including neuroendocrine lesions), 20 (36%) maintained multifocal metaplasia, and the remainder (55%) regressed to focal metaplasia or less worrisome lesions. One patient with dysplasia on index endoscopy developed carcinoma and underwent gastrectomy.

Multivariable logistic regression analysis showed that patients with a high-risk indication for index endoscopy based on clinical outcomes research initiative (CORI) terms (weight loss, iron deficiency anemia, abnormal imaging, gastric ulcer, and epigastric abdominal pain) were more likely to undergo surveillance (OR 2.03, CI 1.20 – 3.45) (Table 1). Those found to be H. pylori positive during the index endoscopy were less likely to follow-up for surveillance (OR = 0.47, CI 0.26 – 0.87).

CONCLUSION:
Surveillance of GIM in a high-risk American population has a significant yield, particularly among those with multifocal involvement. However, in the underserved setting, surveillance compliance is suboptimal. Our findings underscore the need to develop strategies to improve participation and compliance in these important programs.