High absorptivity and turbidity interfere with the UV disinfection of apple cider. Three different configurations of flow-through UV reactors were evaluated to overcome this interference. Two approaches were employed: use of an extremely thin film UV reactor and increasing the turbulence within a UV reactor. Multiple-lamp UV reactors including the thin-film laminar flow "CiderSure (8 lamps) and turbulent flow "Aquionics (12 lamps) and annular single-lamp "UltraDynamics reactor were studied. UV disinfection performance in laminar and turbulent flow reactors was compared by evaluation of UV dose delivery. UV fluence rate (irradiance) distribution was calculated using the multiple point source summation method. E. coli K12 was used as a target bacterium in a bioassay, and the log reduction per one pass was determined for each UV reactor. Finally, the UV decimal reduction dose (D-10) was calculated by dividing the average UV fluence by log bacterial reduction per pass. Variations of the UV decimal dose were observed with various designs of UV systems. The least inactivation of E. coli K12 but the highest UV decimal reduction dose, ranging from 90 to 150 mJ/cm(2) , was observed in the Aquionics UV reactor in apple cider with apparent absorption coefficient ( a ) of 5.7 mm(-1) . The lower value of UV decimal reduction dose of 7.3-7.8 mJ/cm(2) was required for inactivation of E. coli K12 in malate buffer and apple juice in the annular single-lamp UltraDynamics reactor. However, the decimal reduction dose for E. coli K12 in apple cider was significantly higher, about 20.4 mJ/cm(2). Similar UV decimal reduction doses from 25.1 to 18.8 mJ/cm(2) for inactivation of E. coli K12 were observed in the thin-film 'CiderSure UV reactor in apple cider with identical absorption coefficient. Mathematical modeling of UV irradiance can improve the evaluation of UV dose delivery and distribution within the reactors.