Excess 231 Pa and 230 Th (231 Pa xs and 230 Th xs) can be used to reconstruct past oceanic sedimenta-tion (230 Th-normalized flux) and circulation changes ((231 Pa/ 230 Th) xs,0 , hereafter Pa/Th). These quantities are determined by computing the detrital and authigenic contributions from bulk sediment measurement. The method relies on the use of a chosen constant value of the detrital (238 U/ 232 Th) activity ratio (hereafter (U/ Th) det). In this study, we have extracted the detrital fraction of the sediments from North Atlantic deep-sea core SU90-08 (43803 0 1N, 30802 0 5W, 3,080 m) and determined its (U/Th) det value over the last 40 ky. We find that (U/Th) det varied significantly through time with a minimum value of 0.4 during the Holocene and a maximum value of 0.7 during the Last Glacial Maximum (LGM). The sensitivity of sedimentary 230 Th-normalized flux and Pa/Th is tested for our study site and for other North Atlantic sites. We show that the sensitivity is highly dependent on the core location and its terrigenous material supply. The 230 Th-normalized flux and Pa/Th signals are very robust in cores with low detrital contributions, whereas they are very sensitive to (U/ Th) det changes in cores with higher detrital contribution. In the latter case, changes in 230 Th-normalized flux and Pa/Th due to the choice of a constant (U/Th) det can largely exceed the uncertainty on the 230 Th-normalized flux and Pa/Th, inducing potential biases in the amplitude and temporal variability of reconstructed sedimentation and ocean circulation changes.