Apple extract powders from three different manufacturers were investigated for their anti-inflammatory activity, their total phenolic content, and their chemical composition. The samples represented two production batches for two products and a single batch of a third. The samples showed similar, but clearly different, anti-inflammatory activities, and had substantially different total phenolic contents, and different chemical compositions. Differences in chemical composition for batches of the same product were significant, although not as great as differences between products. The samples were fractionated into chemical classes. The most active fractions were those that contained epicatechin, catechin with phloridzin and quercetin glycosides, or those that contained procyanidin polymers. It was not possible to link activity to the presence of individual components or combinations of these. If fruit extracts are to be reliably linked to validated health benefits, then the source materials, the extraction processes, and the final composition of such products need to be more clearly defined than at present.