Continuation rates with a levonorgestrel-releasing contraceptive implant (Norplant). A prospective study in Belgium.

Contraceptive protection offered by a method depends on its duration of use, which reflects costs, side effects, and relatives' opinions. This study investigated in Norplant implants users the continuation rates, some of their determinants, and the motives for removals. Since 1988, 612 Norplant implants sets, designed to protect for 5 years, have been inserted. Observing 13,907 months of use, we determined over time the continuation rates and how age, parity, circumstances at insertion (postpartum, postabortum, others), and patronymic origins (surrogate for sociocultural factors) influenced them. Statistics included Kaplan-Meier's method and log rank tests, and uni- and multivariate Cox models. Continuation increased with age and depended on sociocultural factors. Parity exerted influence only in younger women. Median duration of use was 3 years 11 months. Removals before 5 years related almost equally to irregular bleeding, other side effects, and pregnancy wish. The cumulative 5-year failure rate was 1.5%. Unsatisfied users returned earlier, distorting the first results. A literature search showed that implants yield, in the mean, slightly better continuation figures than do intrauterine devices, and clearly higher than those obtained with pills and injectables. To optimize costs and counseling, warnings about the risk of short duration of use in young nullipara, especially if negative sociocultural influences prevail, are recommended. In no category are the implants absolutely to be avoided. Individual and programmatic contraceptive choice should take into account the expected continuation of use.