We have evaluated the effect of delivering nitric oxide using a continuous flow system (CFS) or two commercially available proportional gas injection systems (PGIS), Nodomo (Dräger, Lübeck, Germany) and Pulmonox-Mini (Messer Griesheim Austria, Gumpoldskirchen, Austria) on measured and simulated concentrations of nitric oxide. Nitric oxide concentration was measured in a bench test at five sites in the inspiratory breathing system during volume- or pressure-controlled ventilation and mathematically simulated using a mixing chamber model. For a target concentration of 10 parts per million (ppm) at the "Y" piece, simulated nitric oxide concentrations were 1.9-139 ppm for CFS, 0.3-22 ppm for the Nodomo and 0.0-31 ppm for the Pulmonox-Mini near the nitric oxide administration site. However, peak concentrations decreased rapidly along the inspiratory system. Measured and simulated variations depended on the nitric oxide delivery system, site of measurement and tidal volume. Measured variations were four times smaller in the Nodomo than in the Pulmonox-Mini and CFS. As inappropriate mixing may occur even with PGIS, nitric oxide should probably not be administered near the "Y" piece.