A community-based longitudinal study conducted in rural Bangladesh investigated the association between nutritional status, cell-mediated immune status and acute upper respiratory infections (URI). A total of 696 children aged 0-59 months was followed prospectively for 1 y yielding 183,865 child-days' observation. Trained field workers visited each child every 4th d and collected morbidity data on symptoms suggesting URI (cough, fever, nasal discharge) for the preceding 3 d by recall. On the day of visit they examined each child reporting cough and/or fever to record the temperature, presence of nasal discharge, rate of respiration and presence of chest indrawing. Anthropometry for all children was conducted monthly. Cell-mediated immune competence was assessed by a multiple antigen skin test at baseline and thereafter every 3 months. The incidence of URI was 5.3 episodes per child-year observed. Approximately three-quarters of the study children were below -2 Z-score weight for age and height for age, and a quarter below -2 Z-score weight for height. During different test periods 9-21% of the study children did not respond to any of the test antigens. In a regression model children < -2 Z-score for weight for height had 16% [odds ratio (OR) 1.16, 95% confidence interval (CI) 1.03-1.31, p = 0.01] higher risk of developing URI. Anergic children had 20% higher risk (OR 1.20, CI 1.05-1.38, p = 0.009) of URI than immunocompetent children. The study demonstrated that wasting and depressed cell-mediated immunity (CMI), but not stunting, were associated with the incidence of URI among rural Bangladeshi children.