Objectives: This paper reports on the replacement risk of different tr
eatment modalities for Class II amalgam restorations in a clinical tri
al of 15 years' duration, Methods: The performance of 1117 conventiona
l Class II amalgam restorations in a controlled, longitudinal study we
re analysed using logistic regression with a random component. Primary
variables regarding replacement risk were the treatment modality (cav
ity wall treatments) and alloy (conventional versus high copper). Seco
ndly, the operator, type of tooth and type of restoration (MO/DO vs MO
D) were considered. Results: Over 15 years, 17% of the restorations we
re replaced (true failures). The application of copalite varnish or si
lver suspension and the type of alloy did not reduce the replacement r
isk. Reduced risks were observed by providing a 90 degrees cavosurface
angle combined with a cavity wail finish, The operator and the type o
f restoration determine replacement risk to a significant extent. Conc
lusions: Additional treatment modalities do not necessarily reduce rep
lacement risks of Class II amalgam restorations within 15 years, while
clinical variables affect the risk of replacement to a certain degree
. (C) 1998 Elsevier Science Ltd. All rights reserved.