Recent life cycle assessments (LCAs) of buildings highlight the importance of global warming potential from construction materials, in particular in energy-efficient buildings. It is therefore important to address the influence of methodological choices related to materials on LCA results. This paper focuses on scenarios for the maintenance and replacement of building elements. Methods: A literature review is carried out to summarize the state of the art regarding scenarios for maintenance and replacement in building LCA, their influence on LCA results and related methodological issues. Additionally, a case study is carried out to investigate whether assumptions about service lives in LCA could significantly influence the recommended design for a building's roof, using a Monte Carlo analysis considering service lives as stochastic variables. Results: The literature review reveals a broad range of impacts from maintenance and replacement in case studies. There is therefore no consensus about the relative impact of these processes. These differences can be partly explained by differences in scope (e.g. what elements are considered to be replaced and what kinds of processes are included), in methods for service life estimation and in future scenarios for the production and recycling of materials. Relative impacts from maintenance and replacement seem to be highest for energy efficient buildings with a long service life, and for elements such as carpets, paint, insulation, doors and windows. The case study of roofing materials exemplifies a case where assumptions about service lives could influence design decisions. Both the ranking of alternatives and the relative significance of maintenance and replacement processes depend on assumptions about service lives. An asphalt roof cover is preferred when considering only initial installation, but a clay tile roof cover is preferred over asphalt in roughly two thirds of the cases when considering maintenance and replacement. Metal roofs almost always had a poorer environmental performance under the assumptions considered. Conclusions: Results from the case study are compared with previous studies of maintenance and replacement processes, and methodological issues deserving further consideration are highlighted. In particular, the case study is used to discuss the issue of whether a modelling based on independent service lives for various building elements accurately reflects industrial practices. Moreover, the relevance of including maintenance and replacement in regulations and climate declarations for buildings is discussed.