The carbon isotopic signal of the C37-alkenone, a taxon-specific biomarker for haptophyte algae, has been used in various paleoceanographic studies as a proxy for ancient surface water CO2 concentration ([CO2aq]). However, a number of recent culture, field and sediment studies imply that the carbon isotopic fractionation (ep) of alkenones is controlled predominantly by physiological processes and environmental factors other than the ambient CO2 concentration (i.e., growth rate, nutrient availability, light intensity, active carbon uptake, bicarbonate utilisation). The environmental conditions controlling phytoplankton growth are likely to vary strongly with oceanographic setting. Culture experiments can not perfectly recreate natural growth conditions and physical processes which affect the carbon isotopic signal in the field and its preservation in the sediment. Consequently, the use of the carbon isotopic record of alkenones as a reliable paleoceonographic proxy also requires sediment-based studies covering a broad range of different oceanic regimes for the past and modern ocean. Here, we present the first basin-wide comparison of alkenone ep values from sediments of the Last Glacial Maximum (LGM) and the latest Holocene. Different oceanographic regions from the equatorial and South Atlantic Ocean were examined. Generally, alkenone ep is lower during the LGM compared to the Holocene. Considering present understanding of LGM-Holocene changes in surface water conditions, the observed glacial/interglacial difference in ep indicates that different effects controlled the isotopic fractionation in alkenone producing algae depending on the regional setting. In upwelling regions, the variations in ep probably reflect a glacial increase in haptophyte productivity controlled by the availability of surface water nutrient concentrations. By contrast, in oligotrophic areas slightly lower nutrient content was available during LGM. Here, the observed ep difference can be explained partly with an assumed glacial decrease in surface water [CO2aq]. However, it can not be ruled out that changes in haptophyte productivity also affected the ep signal to some extent. This study clearly demonstrates that a reliable reconstruction of [CO2aq] on the basis of the isotopic composition of alkenones is not feasible without a detailed knowledge of ancient haptophyte growth conditions.