The long-term performance of bioprostheses was evaluated in women 35 years of age or less to determine the influence of pregnancy on structural valve deterioration. Between 1972 and 1992, 237 female patients received 255 biological prostheses. Of the total operations, 53 were performed in patients who experienced pregnancy (P) and 202 in patients who were never pregnant (nonpregnant [NP]). The mean age of the P group was 23.0 +/- 5.8 years (standard deviation) (12 to 34 years) and of the NP group it was 27.1 +/- 6.3 years (8 to 35 years) (p < 0.05). The mean follow-up for the NP group was 6.8 years and for the P group it was 7.9 years. The late mortality was 2.26%/patient-year overall, 2.71%/patient-year for the NP group and 0.89%/patient-year for the P group (p = not significant [NS]). The P group of 52 patients had 94 pregnancies: 70 deliveries (74.5%) and 24 abortions (25.5%) (therapeutic, 14 [15%]). There were a total of 143 valve-related complications (P, 35; NP, 108); the majority for structural valve deterioration (SVD) 43% (109 patients), P 51% (27 patients) and NP 41% (82 patients) (p = NS). The valve-related reoperation rate paralleled the SVD rate at 42% (107 patients), P 51% (27 patients) and NP 40% (80 patients) (p = NS). The overall reoperative mortality rate was 6.0%. The interval from initial implant to reoperation was 99.6 +/- 3.6 months (p = NS groups P and NP).(ABSTRACT TRUNCATED AT 250 WORDS)