The durability of silicone gel-filled breast implants is of concern, but there are few epidemiological studies on this issue. To date, most of the relevant findings are derived from studies of explantation, which suffer from bias by including women with symptoms or concerns about their implants. As part of a long-term magnetic resonance imaging study of the incidence of rupture, this study involved 271 women with 533 cosmetic breast implants who were randomly selected from among women who underwent cosmetic breast implantation from 1973 through 1997 at one public and three private plastic-surgery clinics in Denmark. The prevalence of rupture was determined from the first magnetic resonance screening. The images were evaluated by four independent readers, using a standardized, validated form. The outcomes under study were rupture, possible rupture, and intact implant. Ruptures were categorized as intracapsular or extracapsular. Overall, 26 percent of implants in 36 percent of the women examined were found to be ruptured, and an additional 6 percent were possibly ruptured. Of the ruptured implants, 22 percent were extracapsular. In multiple regression analyses, age of implant was significantly associated with rupture among second- and third-generation implants, with a 12-fold increased prevalence odds ratio for rupture of implants that were between 16 and 20 years of age, compared with implants between 3 and 5 years of age. Surgitek implants (Medical Engineering Corporation, Racine, Wis.) had a significantly increased prevalence odds ratio of 2.6 for rupture, compared with the reference implants. No significant association was found with the position (subglandular or submuscular) or the type of implant (single- or double-lumen). Extracapsular ruptures were significantly associated with a history of closed capsulotomy (p = 0.001). In the future, the authors plan to examine the women in their cohort with a second magnetic resonance imaging scan to establish the incidence of rupture, a parameter unknown to date in the literature, and to further characterize those factors associated with the actual risk of rupture.