A 20 micrograms/24 hours levonorgestrel-releasing intrauterine contraceptive device (LNG-IUD) has, in an open randomized multicenter study, been compared to the copper-releasing device Nova T during 5 years of use; 1821 women had the LNG-IUD and 937 women had the Nova T inserted. The 5-year cumulative gross pregnancy rate was 5.9% for Nova T and 0.5% for the LNG-IUD. The local effect of levonorgestrel in the uterine cavity causes reduction of menstrual blood loss and development of oligo-amenorrhea, and the termination rates because of heavy and/or prolonged menstrual flow were significantly (P < 0.001) lower among LNG-IUD users. However, the gross termination rate because of amenorrhea was 6.0 in this group. Hb increased during use of the LNG-IUD and decreased during Nova T use, and the difference between the devices was statistically significant. The incidence of pelvic inflammatory disease (PID) was low in LNG-IUD users regardless of age, whereas in the Nova T group, the PID rate was increased among the youngest women, which makes the difference between the devices significant (P < 0.01). The gross termination rate for reasons considered to be hormonal was 12.1 in the LNG-IUD group compared to 2.0 in the Nova T group (P < 0.001). The LNG-IUD is a long-acting reversible contraceptive method for 5 years with a pregnancy rate comparable to female sterilization. The incidence of PID is low even in young women. In addition, the positive health effects such as decreased menstrual blood loss and increased blood haemoglobin concentration make the LNG-IUD a favorable combination of hormonal and intrauterine contraception.