Data on the levels of bacteria and the amounts of food consumed in food-borne outbreaks provides an excellent opportunity to study the effects of exposure to Listeria monocytogenes. Between June 1998 and April 1999, an outbreak caused by L. monocytogenes serotype 3a in butter occurred in Finland. The majority of the cases were immunocompromised and hospitalized at the Helsinki University Central Hospital (HUCH), where 7-g butter packages produced by a dairy plant were used as the only butter brand. The butter had also been sold to 10 other central hospitals as well as to the retail market. Based on the data on hospital stay, butter consumption and the qualitative and quantitative analyses of L. monocytogenes in butter, the attack rates and exposure were estimated. Incubation studies on the naturally contaminated small butter packages showed that the levels found in the packages at the time of detection of the outbreak could reliably be used for these estimations. However, the levels of L. monocytogenes in 500-g packages increased. The attack rate among HUCH patients varied from 70 to 117 cases per 1000 patients at risk, depending on which estimate of the contamination level of butter (100-60%) was used. The highest single dose (7.7 x 10(4) CFU in one meal) could have been sufficient to cause the listeriosis cases at HUCH. However, this data also supports another hypothesis, according to which these listeriosis cases were caused by a prolonged daily consumption of contaminated butter during the hospital stay. The estimated daily dose, based on the hospital kitchen data or the highest detected level in a wholesale sample (11,000 CFU/g), would have varied from 1.4 x 10(1) to 2.2 x 10(3) CFU/day or from 2.2 x 10(4) to 3.1 x 10(5) CFU/day, respectively. The choice of the hypothesis has a crucial impact on the interpretation of this data for the dose-response estimations as well as for the discussion on Food Safety Objectives. Due to the susceptibility of hospital patients, special care must be taken in order to avoid even low levels of L. monocytogenes in food served.