Primary cancer of the head and neck was diagnosed in 676 cases. Systematic endoscopic screening of the upper gastrointestinal tract (including iodine staining of the esophagus) was performed within 1 month of diagnosing head and neck cancer in order to determine the possible presence of a second malignancy. We divided cases in which esophageal cancer was detected into a group in which the initially diagnosed head and neck cancer had been treated curatively and those in which the treatment had been non-curative. A pathological diagnosis of primary esophageal cancer was obtained in 37 cases (5.5%). The incidence of esophageal cancer associated with cancer of the oral cavity or pharynx was significantly higher ( 1 0.9%) than in cases with an associated cancer of the larynx or paranasal sinus (1.9%, p<0.05). All but one of the 37 cases detected by screening were asymptomatic and had superficial carcinoma. Of these lesions, 19 Were mucosal carcinomas, 17 submucosal carcinomas and only one an advanced cancer. Twenty-four of these 37 cases had died. The cause of death was head and neck cancer in 17, esophageal cancer in t w o and other causes in five. The 5-year survival rate of cases in whom the initial head and neck cancer had been treated curatively was 46.7%.The risk of synchronous esophageal cancer associated with cancer of the head and neck is high. Early detection of esophageal cancer at the time of diagnosing cancer of the head and neck and treatment at an early stage is a new strategy for improving the long-term outcome of cases with head and neck cancers associated with esophageal cancer. (Dig Endosc 1998 ; 10 : 110-1 15)