Abstract Contamination from untreated wastewater leakage and related bacterial contamination poses a threat to drinking water quality. However, a quantification of the magnitude of leakage is difficult. The objective of this work is to provide a highly sensitive methodology for the estimation of the mass of untreated wastewater entering karst aquifers with rapid recharge. For this purpose a balance approach is adapted. It is based on the mass flow of caffeine in spring water, the load of caffeine in untreated wastewater and the daily water consumption per person in a spring catchment area. Caffeine is a source-specific indicator for wastewater, consumed and discharged in quantities allowing detection in a karst spring. The methodology was applied to estimate the amount of leaking and infiltrating wastewater to a well investigated karst aquifer on a daily basis. The calculated mean volume of untreated wastewater entering the aquifer was found to be 2.2 ± 0.5 m 3 d −1 (undiluted wastewater). It corresponds to approximately 0.4% of the total amount of wastewater within the spring catchment.