The aim of our study was to evaluate the leakage of four cements (Cavit, IR
M, TERM, and Fermit) using a two-compartment model system and Streptococcus
sanguis as bacterial marker. Access cavities in premolars were filled with
cement and the teeth immersed in culture medium in the model system. Half
of the teeth were thermocycled on day 2, Bacterial percolation into the upp
er compartment was measured at regular intervals (days 2, 7, 14, and 21), C
ement thickness was measured at the end of the study. In the nonthermocycle
d group, Cavit was more leakproof than the other cements at day 2 (p = 0.01
1), than TERM and IRM at day 7 (p = 0.043), Fermit was more leakproof than
IRM at day 7 (p = 0.043), In the thermocycled group, Cavit was more leakpro
of than the other cements at day 7 (p = 0.041), Thermocycling did not signi
ficantly affect leakage. Cement thickness averaged 4.1 mm and did not signi
ficantly affect leakage. These results should be considered when using ceme
nts as temporary fillings.