Geochemical and micropaleontological analyses of Labrador Sea sediments allowed us to intercalibrate indicators of paleoproductivity and paleofluxes of carbon and to reconstitute biogeochemical fluxes, notably in relation to deep sea circulation changes (Western Boundary Undercurrent and North Atlantic Deep Water). Linear or logarithmic relations are observed between the subrecent fluxes of paleoproductivity indicators (coccoliths, diatoms, dinocysts, Th-230), the C-13 content of left-coiling or sinistral Neogloboquadrina pachyderma, the modern surface water carbon production, and the rates of organic and inorganic carbon (C(org), C(inorg)) burial. All indicators increase by a factor of 10(2) to 10(3) in response to the evolution of paleosea-surface conditions since the last glacial maximum. Correlations are observed between authigenic C(inorg), Th-230, coccoliths, and C-13 content (left-coiling or sinistral Neogloboquadrina pachyderma), and also between authigenic C(org), dinocysts, uranium, and sulfur. Relative C(inorg)/C(org) burial rates in the deep sediments ( > 3000 m) allow us to calculate a net CO2 production in the basin. The net CO2 flux was slightly negative between approximately 20 and approximately 15 ka BP. It reached a first maximum at about 9 ka BP (approximately 30 mumol . cm-2 . a-1 CO2) followed by a decrease at approximately 8 ka BP (approximately 20 mumol . cm-2 . a-1). Since about 7 ka BP, it stabilized at - 35 mumol . cm-2 . a-1. The minimum observed at about 8 ka BP corresponds to a significant rate of C(org) burial accompanied with a maximum flux of redox-sensitive indicators (uranium, sulfur). This event does not correspond to any change in sea-surface temperature or salinity, nor in deep circulation, but may be related to the reorganization of the atmospheric circulation during a major breakup stage of the Laurentide ice sheet. The close correlation between the change in CO2 production in the Labrador Sea with variations of atmospheric CO2 concentrations during the last 20 ka BP suggests that the production of biogenic carbonates in subpolar basins plays a role in the global carbon cycle on the scale of glacial-interglacial climate changes.