Breast implants from 300 consecutive patients during the 35-month period from February 1, 1991 to January 1, 1994 were examined. Of these, 214 (71.3%) patients had disruption (frank rupture or severe silicone bleed or both) of one or both implants. Of the 592 implants removed, 376 (63.5%) had disruption. We found virtually no difference in disruption rates between those patients relating symptoms to their implants and those who did not (71.8% vs. 70.9%). However, disruption was directly related to time since implantation. We observed 103 of 142 patients relating various symptoms to their silicone implants. About half (52) reported improvement after implant removal. We found no direct correlation between disruption and repeated closed capsulotomy. Of 505 implants, 69 of 92 implants having had one or more closed capsulotomies were disrupted (75%). However, 245 of 413 implants never having had a closed capsulotomy also showed disruption (59.3%). We found mammography to be of very limited value in assessing disruption. Our study showed a low 16.2% sensitivity rate. The same was true with magnetic resonance imaging interpretation, especially in the diagnosis of severe silicone bleed. Although our series was quite small, magnetic resonance imaging interpretation was found to be an unreliable tool in our study. A Kaplan-Meier survival curve for mammary implants based on this study has proven very helpful and effective in communicating with patients and has served as a guideline in predicting the presence or absence of implant disruption. We have found and predict that most implants have lost or will lose the integrity of the silicone shell between 8 and 14 years, leaving free silicone, intracapsular or extracapsular, in the breast. On the basis of this information, O. Gordon Robinson recommends for his patients the removal of all gel-filled implants, preferably before 8 years from implantation.