Breeding for resistance is an efficient strategy to manage wheat leaf rust caused by Puccinia triticina f. sp. tritici. However, a prerequisite for the directed use of Lr genes in breeding and the detection of new races virulent to these Lr genes is a detailed knowledge on Lr genes present in wheat cultivars. Therefore, respective molecular markers for 18 Lr genes were tested for specificity and used to determine Lr genes in 115 wheat cultivars. Results obtained were compared to available pedigree data. Using respective molecular markers, genes Lr1, Lr10, Lr26, Lr34 and Lr37 were detected, but data were not always in accordance with pedigree data. However, leaf rust scoring data of field trials confirmed the reliability of DNA markers. These reliable marker data facilitated the analyses of the development of virulent leaf rust races from 2002 to 2009 based on released cultivars. A sudden change from low infection rates to susceptibility was observed for Lr1, Lr3, Lr10, Lr13, Lr14, Lr16, Lr26 and Lr37 since 2006. Cultivars carrying several leaf rust resistance genes showed no significant shift to susceptibility except one cultivar which revealed an increasing infection rate at a low level. In summary, it turned out that pedigree data are often not reliable and a detection of Lr genes by diagnostic markers is fundamental to combine Lr genes in cultivars for a durable resistance against leaf rust, and to conduct reliable surveys based on released cultivars, instead of 'Thatcher' NILs.