The faecal contamination of greywater in a local treatment system at Vibyåsen, north of Stockholm, Sweden was quantified using faecal indicator bacteria and chemical biomarkers. Bacterial indicator densities overestimated the faecal load by 100-1000-fold when compared to chemical biomarkers. Based on measured levels of coprostanol, the faecal load was estimated to be 0.04 g person(-1) day(-1). Prevalence of pathogens in the population and the faecal load were used to form the basis of a screening-level quantitative microbial risk assessment (QMRA) that was undertaken for rotavirus, Salmonella typhimurium, Campylobacter jejuni, Giardia lamblia and Cryptosporidium parvum. The different exposure scenarios simulated--direct contact, irrigation of sport fields and groundwater recharge--gave unacceptably high rotavirus risks (0.04 < Pinf < 0.60) despite a low faecal load. The poor reduction of somatic coliphages, which were used as a virus model, in the treatment was one main reason and additional treatment of the greywater is suggested. Somatic coliphages can under extreme circumstances replicate in the wastewater treatment system and thereby underestimate the virus reduction. An alternative QMRA method based on faecal enterococci densities estimated similar risks as for rotavirus. Growth conditions for Salmonella in greywater sediments were also investigated and risk modelling based on replication in the system increased the probability of infection from Salmonella 1000-fold, but it was still lower than the risk of a rotavirus infection.