Bias and confounding in longitudinal measures of exhaled monoxides.

The measurement of exhaled nitric oxide and carbon monoxide concentrations is an emerging method of monitoring airway inflammation longitudinally in community-based studies. Inhaled concentrations of these monoxides influence exhaled concentrations. Little is known about the degree to which inhaled concentrations distort temporal trends in, or estimated effects of air pollutants on, exhaled monoxides. We sought to evaluate whether estimated effects of air pollutants on exhaled monoxides are distorted by trends in indoor and outdoor monoxides, and to characterize determinants of exhaled monoxide concentrations among residents of public housing. In a panel study, 42 residents of public housing provided over 1000 exhaled breath samples. Samples from all subjects were analyzed for nitric oxide; samples from 27 of these subjects were also analyzed for carbon monoxide. The effects of indoor and outdoor monoxide concentrations on exhaled concentrations were quantified. Confounding of associations between particulate matter concentrations and exhaled nitric oxide concentrations was explored. Determinants of exhaled monoxide concentrations among public housing residents are similar to those of other populations. Exhaled monoxide concentrations are more strongly associated with indoor than with outdoor monoxide concentrations. Approximately half of the variability in exhaled monoxide concentrations over time can be explained by changes in indoor monoxide concentrations. Indoor monoxide concentrations can markedly distort both temporal trends in exhaled concentrations as well as estimated effects of particulate matter on exhaled monoxides. Prior estimated effects of particulate matter on exhaled nitric oxide concentrations may have been confounded by nitric concentrations indoors at the time of exhaled air collection. To prevent distortions of longitudinal trends in airway inflammation and estimated health effects of air pollutants, inspiratory scrubber use is necessary but not sufficient to remove the confounding effect of indoor monoxides, and analyses should adjust exhaled monoxide concentrations for concentrations indoors.