In the setting of mechanical ventilation, recent studies have cast doubt on the ability of nebulizer systems to deliver adequate amounts of medication. We therefore studied ventilator-related and nebulizer-related factors that could potentially affect the amount of aerosol inhaled by an intubated subject. Utilizing two separate protocols, we used a bench model of a ventilator circuit, radiolabeled (technetium pertechnetate, 99mTc) saline droplets and a filter technique to measure the percentage of radioaerosol delivered. First, we compared four commercially available jet nebulizers and found that there were significant differences in rate of aerosol production between systems, ranging from 3 to 37%. Delivered aerosol was measured at different ventilator settings, and it was found that the duty cycle can potentially influence output by sevenfold. Some nebulizers were also sensitive to changes in the initial volume of solution placed in the nebulizer. The inclusion of a humidification device significantly reduced output by a mean of 41 +/- 3.5%, but it did not affect particle distribution. Endotracheal tube diameter was not an important variable. Then, with the effects of the above variables established, a separate series of experiments was performed to test whether the use of different radiolabeling compounds can confound the measurement of inhaled drugs. Two nebulizers (AeroTech II and Twin Jet) and pentamidine as the test drug were studied with fixed ventilator settings, treatment time, endotracheal tube size, and volume fill. No humidification was used. The nebulizer solution was labeled with 99mTc, which was bound to either human serum albumin (HSA) or sulfur colloid (SC).(ABSTRACT TRUNCATED AT 250 WORDS)