Background: Pneumonia is a major cause of death in patients with cerebral infarction. We assessed morbidity associated with pneumonia in 276 patients 65 years of age or older who were admitted to a long-term care facility. Furthermore, we studied the swallowing reflex during the day and at night and monitored the occurrence of silent aspiration during sleep. Objectives: To examine the possible relationship between the location of cerebral hemispheric infarctions and the incidence of pneumonia and to evaluate the role of silent aspiration in the development of pneumonia. Methods: The incidence of pneumonia was analyzed in 4 groups of patients who were assigned to a group on the basis of the following computed tomographic findings: no infarct (group A); 1 or more unilateral basal ganglia infarcts (group B); bilateral basal ganglia infarcts (group C); and 1 or more cerebral hemispheric infarcts outside the basal ganglia (group D). Criteria for diagnosis of pneumonia were (1) a new pulmonary infiltrate seen on a chest radiograph and (2) 1 or more of the following features: cough, temperature greater than 37.8°C, or subjective dyspnea. Before the study, the patients with stroke were followed up for more than 1 year after their ictus and were monitored to determine if they sustained affecting cerebral hemispheric structures. The average duration of observation for incidence of pneumonia was 22 months. To study the swallowing reflex and to monitor for the occurrence of silent aspiration during sleep, 15 of the patients who were confined to bed or chair were randomly selected from each of groups A through C. The swallowing reflex was examined at both 1 PM and 1 AM and was evaluated according to latency of response, which was timed from the injection of 1 mL of distilled water into the pharynx through a nasal catheter to the onset of swallowing. The incidence of silent aspiration during sleep was examined using indium-111 chloride as a radioactive tracer attached to the teeth, and scanning of the thorax was performed the next morning. Results: The incidence of pneumonia was 2.12 times higher in the patients of group B (27.4%; P<.01) and 3.64 times higher in the patients of group C (47.0%; P<.001) than in the patients of group A (12.9%). The latency of response was longer in the patients of groups B (P<.05) and C (P<.001) than in those of group A at 1 AM. The percentage of positive scans was also higher in the patients of groups B (P<.01 ) and C (P<.001 ) than in those of group A. Conclusion: Basal ganglia strokes might predispose these patients to develop pneumonia owing to frequent aspiration during sleep. Arch Intern Med. 1997;157:321-324