We analyzed suicide rates in Sweden 1952-1992 with the main purpose of following up the previously observed increasing suicide rate in young men, and applying age-period-cohort (APC) analyses to the trends in suicide mortality. APC analyses were performed by a graphical method and by multivariate log-linear regression. The suicide rate among 20-40 year-old men increased throughout the 1950s and 60s, but the increase has levelled off since the middle of the 1970s, and in some narrow age groups possibly even reversed. The suicide rate among men over 45 years has declined throughout the period. The suicide rate among women has remained more stable. APC analyses did not give clear evidence for a specific cohort or period effect, although addition of a cohort term in the analyses of men slightly improved the fit of the model. A longer follow-up of younger birth cohorts is needed to see whether the changes in male suicide rates will remain as a cohort effect.