This in vitro study investigated the effects of a thin layer of flowable composite or compomer on microleakage occurring in Class II packable and nanofilled composite restorations that extend apical to the cemento-enamel junction (CEJ). The current study also investigated any differences in microleakage that occur between restorations light-cured using a light-emitting diode (LED) and a quartz tungsten halogen (QTH) light-curing unit. Standardized Class II "slot" cavity preparations were prepared on the mesial and distal surfaces of 72 extracted, unrestored, non-carious premolars (n = 144). The gingival margins were placed 1 mm apical to the CEJ. The teeth were divided into two groups (n = 72) and restored with a "packable composite" (Filtek P60) or a "nano-filled composite" (Universal Filtek Supreme XT) with or without flowable composite (Flowable Filtek Supreme XT) or flowable composite (Dyract Flow) as gingival liners placed with thicknesses of 1.0 mm. Each increment was cured for 20 seconds before adding the next. One-half of the samples in each group was cured with QTH (Coltolux 75) and the other half with LED (Coltolux LED) LCUs. After a two-week incubation period at 37 degrees C, the specimens were thermocycled (5 degrees C-55 degrees C x 1500), immersed in 0.5% basic fuchsin dye for 24 hours, sectioned and the microleakage was then evaluated at the gingival margin by two examiners using a 0-3 score scale. Within the current study, when flowable liners were used, both the packable (Filtek P60) and nanofilled (Filtek Supreme XT Universal Restorative) composite materials had significantly less microleakage than when flowable liners were not used (p < 0.05). Both flowable liners (Flowable Filtek Supreme XT and Dyract Flow) resulted in a significant reduction of the microleakage occurring under both types of composite materials at the gingival floors ( p < 0.05), but there was no significant difference between them. The choice of light curing technology (LED vs QTH) had no significant effect on the amount of microleakage observed.