SUMMARY Objective To compare the ability of two calcium-releasing restorative materials to inhibit root dentin demineralization in an artificial caries model. Methods and Materials Preparations were made at the cementum–enamel junction of extracted human molars (40, n=10/material) and restored with two calcium-releasing materials (Experimental composite, Pulpdent Corporation and Cention N, Ivoclar Vivadent), a resin composite (Filtek Supreme Ultra, 3M Oral Care), and a resin-modified glass ionomer (RMGI) (Fuji II LC, GC). All materials (other than the RMGI) were used with an adhesive (Scotchbond Universal Adhesive, 3M Oral Care) in the self-etch mode, which was light cured for 10 seconds. All restorative materials were light cured in 2-mm increments for 20 seconds and then finished with a polishing disc. Teeth were incubated (37°C) for 24 hours in water. An acid-resistant varnish was painted onto the teeth around the restoration, leaving a 2-mm border of uncovered tooth. A demineralization solution composed of 0.1 M lactic acid, 3 mM Ca3(PO4)2, 0.1% thymol, and NaOH (to adjust pH=4.5), and a remineralization solution composed of 1.5 mM Ca, 0.9 mM P, and 20 mM Tris(hydroxymethyl)–aminomethane (pH=7.0) were prepared. Specimens were placed in the demineralization solution for 4 hours, followed by the remineralization solution for 20 hours and cycled daily for 30 days. The specimens were embedded, sectioned into 100-μm sections, and the interface between the restorative material and root dentin was viewed with polarized light microscopy. A line was drawn parallel with the zone of demineralization for each tooth. The area of “inhibition” (defined as the area external to the line) or “wall lesion” (defined as the area internal to the line) was measured with image evaluation software. Areas of inhibition were measured as positive values, and areas of wall lesions were measured as negative areas. Results A one-way analysis of variance (ANOVA) found significant differences between materials for “inhibition/wall lesion” areas in root dentin (p<0.001). Tukey post hoc analysis ranked materials (μm2, mean ±SD): Fuji II LC (5412±2754) > Cention N (2768±1576) and experimental composite (1484±1585) > Filtek Supreme Ultra (−1119±1029). Conclusion The experimental composite and Cention N materials (used with an adhesive) showed net areas of inhibition greater than a reference resin composite, albeit at a lower level than a reference RMGI material (used with no adhesive).