Anomalously enhanced NO₂ concentrations are sometimes observed in the polar winter upper atmosphere. The enhancements over Antarctica from May to August 2003 were probably due to auroral electron precipitation, producing high amounts of NO in the upper mesosphere that were converted to NO₂ during downward transport to the stratosphere. Another enhancement was detected in the Arctic middle stratosphere in October–November 2003, due this time to energetic solar proton precipitation. This enhancement was quickly followed by a new Arctic NO₂ enhancement produced by auroral electrons in November 2003. Finally, a last enhancement was detected in the lower mesosphere from January to April 2004. Although it was proposed that this enhancement could also be due to auroral electrons, uncertainties remained concerning the absolute value of the NO₂ enhancement and its spatial coverage. We propose here a new analysis of the Global Ozone Monitoring by Occultation of Stars (GOMOS) nighttime measurements of the NO₂ enhancements. Instead of using daily zonally averaged data as done previously, we consider only the profiles containing the maximum values of the NO₂ enhancement. Unlike all the previous enhancements, the NO₂ content of the January 2004 appears to be longitudinally and latitudinally dependent inside the polar circle. The enhancement starts on 17 January 2004, with mixing ratios of up to a ppmv at altitudes above 60 km on 21 January. The enhancement looks like a “hot spot” above the polar cap. Then the enhancement spreads while it propagates downward into the stratosphere. It is accompanied by perfectly coincident strong ozone depletion; in particular, ozone is almost totally destroyed in mid-February at about 50 km. The vertical extent and horizontal spread of this NO₂ enhancement strongly differ from the November 2003 enhancement attributed to auroral electron precipitation. The possible origins of this unusual pattern are discussed.