Background: Gastric cancer may be suspected with otherwise unexplained positive faecal occult blood testing.
Aims: To assess the frequency of gastric cancer following positive faecal occult blood testing and negative colonoscopy.
Subjects: Age 40-74 cohort at first screening (1985-2001) with (a) faecal occult blood testing- (83,489), (b) faecal occult blood testing +/colonoscopy+ (2025), or faecal occult blood testing+/colonoscopy- (3555).
Methods: Gastric cancer incidence in faecal occult blood testing subsets, compared with expected standardized incidence rates.
Results: Gastric cancer risk was increased (standardized incidence rate=146.7; 95% confidence interval: 105.8-203.4) in faecal occult blood testing+/colonoscopy- subjects. A four-fold excess incidence occurred during first year (observed cases=10, standardized incidence rate=408.3; 95% confidence interval: 219.7-758.8), irrespective of faecal occult blood testing type (guaiac, immunological). No excess risk occurred in faecal occult blood testing- (observed cases=53, standardized incidence rate=91.2; 95% confidence interval: 84.1-98.8) or in faecal occult blood testing+/colonoscopy+ subjects (observed cases=2, standardized incidence rate=101.9; 95% confidence interval: 25.5-407.4). Assuming a 100% 3-year study sensitivity for gastric cancer, faecal occult blood testing positive predictive value would be 0.4% (40-74 years) or 0.7% (> or =60 years) in faecal occult blood testing+/colonoscopy- subjects.
Conclusions: Data suggest an association of faecal occult blood testing+/colonoscopy- and excess gastric cancer incidence in the following year. Due to low faecal occult blood testing+ positive predictive value, routine upper digestive tract endoscopy in these subjects is questionable.