Atmospheric blocking is an important source of low-frequency variability. We apply a blocking detection and tracking method to ERA-40 data for the Atlantic-European region to assess linkages to extreme events, that is, cold and dry spells in the extended winter season (November-April). The method is feature-oriented, identifies 500-hPa geopotential height maxima, and connects them with a next-neighbourhood search in time. The analysis reveals a statistically significant decrease of number of blocking events over the period of ERA-40. Winters with an increased number of blocking events are associated with negative temperature anomalies over Central to Eastern Europe and dryer conditions, whereas Southern Europe experiences warmer and wetter conditions during such episodes. Using extreme value statistics we show evidence that cold spells, and to some extent dry spells, are strongly associated with extremes in blocking frequency. The number of cold spell days increases with the duration of blocking events indicating that cold spells need time to establish during blocking events, thus return periods of cold spells are longer than those for blocking events. This behaviour is not found for the relation of dry spells with blocking events. Still, blocking episodes have a higher occurrence of dry spells than climatology.