Rapid climate changes are occurring in the Arctic, with substantial repercussions for arctic ecosystems. It is challenging to assess ecosystem changes in remote polar environments, but one successful approach has entailed monitoring the diets of upper trophic level consumers. Quantitative fatty acid signature analysis (QFASA) and fatty acid carbon isotope (δ13C-FA) patterns were used to assess diets of East Greenland (EG) polar bears (Ursus maritimus) (n = 310) over the past three decades. QFASA-generated diet estimates indicated that, on average, EG bears mainly consumed arctic ringed seals (47.5 ± 2.1%), migratory subarctic harp (30.6 ± 1.5%) and hooded (16.7 ± 1.3%) seals and rarely, if ever, consumed bearded seals, narwhals or walruses. Ringed seal consumption declined by 14%/decade over 28 years (90.1 ± 2.5% in 1984 to 33.9 ± 11.1% in 2011). Hooded seal consumption increased by 9.5%/decade (0.0 ± 0.0% in 1984 to 25.9 ± 9.1% in 2011). This increase may include harp seal, since hooded and harp seal FA signatures were not as well differentiated relative to other prey species. Declining δ13C-FA ratios supported shifts from more nearshore/benthic/ice-associated prey to more offshore/pelagic/open-water-associated prey, consistent with diet estimates. Increased hooded seal and decreased ringed seal consumption occurred during years when the North Atlantic Oscillation (NAO) was lower. Thus, periods with warmer temperatures and less sea ice were associated with more subarctic and less arctic seal species consumption. These changes in the relative abundance, accessibility, or distribution of arctic and subarctic marine mammals may have health consequences for EG polar bears. For example, the diet change resulted in consistently slower temporal declines in adipose levels of legacy persistent organic pollutants, as the subarctic seals have higher contaminant burdens than arctic seals. Overall, considerable changes are occurring in the EG marine ecosystem, with consequences for contaminant dynamics.