OBJECTIVES: Aim of this study was to assess the marginal adaptation (MA) of Class II IPS Empress^® ceramic inlays placed with various luting materials in vitro. METHODS: Thirty approximal cavities were prepared in extracted human molars with diamond burs, the proximalmargins of the cavities ended 1mm beyond the CEJ. Prepared teeth were randomly assigned to five groups of six teeth each. Inlays were luted using the composite based material Variolink^® II (Ivoclar^® Vivadent^®) and Syntac^® Classic adhesive system [Group A], the composite resin cement PANAVIA^® 21 (KURARAY) [B], the self-etching composite resin Rely^™X Unicem Aplicap (ESPE) [C], the zinc oxide resin cement Auracem^® (DMG Dental) [D] or an experimental self-etching material based on composite resin and zinc oxide Compophat^® (Merz Dental) [E]. Inlays of Groups A to D were conditioned with hydrofluoric acid and silanized, for Group E, inlays were only air-abraded. MA was evaluated by SEM-analysis after placement, thermocycling [TC] and mechanical loading [ML]. RESULTS: Groups A, B and C revealed high percentages of perfect marginal seal between Inlays and luting agents in over 95% of analyzed margins before and after loading. However, in Groups D and E more than 55% of the interface between inlays and luting agents showed gap formation after loading. Mean percentage of marginal gaps were:

	A		B		C		D		E
	enamel	dentin	enamel	dentin	enamel	dentin	enamel	dentin	enamel	dentin
After placement	3.0%	14.9 %	4.6%	7.3%	36.5%	21.7%	7.2%	9.4%	9%	2%
After TC & ML	23,4%	29,6%	12.4%	26.8%	43.7%	22.5%	45.9%	27.8%	45%	15,4%

Conclusion: Load resistant, stable bonding between ceramic restorations and luting agent is not achievable with both Auracem^® and the experimental material Compophat^®. None of the luting systems tested under the conditions of this in vitro study provided perfect marginal adaptation of ceramic inlay restorations after loading.