Functional foods offer the possibility to modulate the absorption of sugars, leading to benefits for diabetics and those with metabolic syndrome. As part of the characterisation of such foods, inhibition of alpha-amylase is used to assess components for their potential ability to modify the post-prandial glycaemic response. Many publications on phenolics as potential inhibitors report widely varying assay conditions leading to variable estimates of inhibition. On this basis, we have optimised the in vitro alpha-amylase inhibition assay and, in particular, we show the importance of removing certain polyphenols after the enzymic reaction when using 3,5-dinitrosalicylic acid since they interfere with this reagent. There was a substantial similar to 5-fold effect on acarbose IC50 values when working just outside optimal conditions. This shows that inappropriate assay conditions, such as excess enzyme, greatly influence IC50 values and could explain some discrepancies in the existing literature. (C) 2015 Elsevier Ltd. All rights reserved.