The carbon isotopic fractionation during photosynthesis (Ɛp) from sedimentary alkenone biomarkers produced by coccolithophores is a widely used proxy for concentrations of past dissolved CO2 in seawater (CO2[aq]). Currently, Ɛp records covering the last 1 Myr exist only for low-latitudes regions, which are characterized by stratified waters and nutrient (NO3− and PO4−) limitation. Higher latitudes are affected by more variable hydrographic conditions which are expected to produce larger changes in factors which are important for Ɛp, such as light, temperature, CO2 and growth rate. Understanding these processes at high latitudes is important in order to derive correct estimates of past CO2 concentrations. Here we present new Ɛp and alkenone-based SST records and a review of previously published paleoceanographic multi-proxy datasets, from sites across a latitudinal transect in the North Atlantic (61°37°N) covering the interval between ∼800 and 400 ka. During this period, the subpolar hydrographic fronts shifted latitudinally, following orbital and sub-orbital climate variability, leading to large zonal and meridional environmental gradients in the North Atlantic. We observe that Ɛp and climate-state relationships (depicted by benthic δ18O) are similar across different latitudes. Ɛp is lower at mid-latitude regions (Sites U1385 and U1313), due to a higher coccolithophore growth rate during colder intervals. In the high-latitude sites, a longitudinal gradient is observed, with higher Ɛp during glacial intervals eastwards (Site 982) compared to interglacials. The opposite is observed on the sites located westwards (Sites U1314 and 984), where Ɛp is higher during interglacials compared to glacials. We suggest that the depressed interglacial Ɛp values at Site 982 is the result of the uninterrupted northward flow of warm Atlantic waters towards the east, which sustained high coccolithophore productivity and growth rates. Combining high-latitude planktonic foraminifera species and calcium carbonate records, we reconstructed the effect of non-CO2 factors on Ɛp.