A planned 21 km bypass (18 km within a tunnel) in Stockholm is expected to reduce ambient air exposure to traffic emissions, but same time tunnel users could be exposed to high concentrations of pollutants. For the health impacts calculations in 2030, the change in annual ambient NOX and PM10 exposure of the general population was modelled in 100 × 100 m(2) grids for Greater Stockholm area. The tunnel exposure was estimated based on calculated annual average NOX concentrations, time spent in tunnel and number of tunnel users. For the general population, we estimate annually 23.7 (95% CI: 17.7-32.3) fewer premature deaths as ambient concentrations are reduced. At the same time, tunnel users will be exposed to NOX levels up to 2000 μg/m(-3). Passing through the whole tunnel two times on working days would correspond to an additional annual NOX exposure of 9.6 μg/m(3). Assuming that there will be ~55,000 vehicles daily each way and 1.3 persons of 30-74 years of age in each vehicle, we estimate the tunnel exposure to result in 20.6 (95% CI: 14.1-25.6) premature deaths annually. If there were more persons per vehicle, or older and vulnerable people travelling, or tunnel dispersion conditions worsen, the adverse effect would become larger.