CO2 samples collected by aircraft near the winter midlatitude and high-latitude Northern Hemispheric tropopause were measured for their stable isotope (13C/12C, 18O/16O) and radioisotope (14C/12C) ratios. The strongly varying CO2 and δ13C(CO2) values spanning 21ppm and 1.1‰, respectively, as well as the low 13C/12C ratio of the source/sink system responsible for these variations (inferred by applying a simple “Keeling relationship”) point to frequent transport of polluted air masses to the tropopause. This hypothesis is supported by the often depleted 14C/12C ratios in CO2, marking contributions of up to 9ppm (14C-free) fossil fuel combustion derived CO2. The oxygen isotope ratio δ18O(CO2) was found to correlate negatively with the CO2 mixing ratio (R ≈ −0.8), which demonstrates that even the δ18O(CO2) data can, as a first approach, be interpreted in terms of a Keeling relation. However, the apparent δ18O(CO2) source/sink signature was found to drop from −(11±3)‰ south of the polar front down to −(27±4)‰ north of it. The low Arctic δ18O(CO2) values can be explained by the assumption that in the wintertime Arctic about double the amount of CO2 isotopically exchanges with 18O-depleted soil water as is net released by the entire biosphere. A vertical δ18O(CO2) gradient of 0.5‰ km−1 was observed above the tropopause. This δ18O(CO2) increase in the stratosphere is most likely due to oxygen isotope exchange between CO2 and electronically excited oxygen O(1D), the isotope composition of which is controlled by that of O3, in the stratosphere known to be strongly enriched in the heavy oxygen isotopes. The typical δ18O(CO2) gradient is assumed to be lower compared to measured because our high-altitude samples were affected by chemically disturbed polar vortex air.