The impacts of rising mean sea levels will be felt most acutely during periods of extreme high sea levels which are caused by the combination of mean sea level, tides and storm surges. In this paper, we examine sea level records from six tide gauges along the German North Sea coastline to determine if changes in high sea levels observed throughout the 20th century and early 21st century were primarily driven by increases in mean sea level (i.e. like what has been observed by other authors in most parts of the world) or whether other factors, such as changes in ocean tides or storm surges also contributed significantly to observed changes in high water in this region. Time-series of annual 80th, 85th, 90th, 95th, 99th, and 99.9th percentiles are derived from the sea level records and trends are assessed using linear regression for the entire time periods for which datasets are available at each site and for the common period from 1953 to 2008. The percentile time-series are subsequently reduced relative to mean sea level and a second set of trends are estimated. At all sites and percentile levels, significant positive trends are evident for the observed sea level data. Once the percentile time-series are reduced relative to mean sea level the remaining trends are still significant at the 1σ-confidence level, with the exception of the 99.9th percentiles since the standard errors are large. Using a non-linear trend analysis, on the long Cuxhaven record, we find that prior to the mid-1950s and from about 1990 onwards, changes in high sea levels were not different from mean sea level changes. However, from the mid-1950s to 1990 changes were significantly different from those observed in mean sea level. Possible reasons for this appear to be due to changes in the amplitudes of several main tidal constituents, which are apparent since the mid-1950s and decadal variability in the storm activity (with strong westerly winds in the North Atlantic region from 1960 to the 1990s). © 2013 Elsevier Ltd.