To collect the ever-increasing yet scattered knowledge on metabolism, multiple pathway databases like the Kyoto Encyclopedia of Genes and Genomes have been created. A complete and accurate description of the metabolic network for human and other organisms is essential to foster new biological discoveries. Previous research has shown, however, that the level of agreement among pathway databases is surprisingly low. We investigated whether the lack of consensus among databases can be explained by an inaccurate representation of the knowledge described in scientific literature. As an example, we focus on the well-known tricarboxylic acid (TCA) cycle and evaluated the description of this pathway as found in a comprehensive selection of 10 human metabolic pathway databases. Remarkably, none of the descriptions given by these databases is entirely correct. Moreover, consensus exists on only 3 reactions. Mistakes in pathway databases might lead to the propagation of incorrect knowledge, misinterpretation of high-throughput molecular data, and poorly designed follow-up experiments. We provide an improved description of the TCA cycle via the community-curated database WikiPathways. We review various initiatives that aim to improve the description of the human metabolic network and discuss the importance of the active involvement of biological experts in these.