The objectives of this report are to evaluate changes in the preterm birth rate in Sweden 1973-2001. Furthermore, describe the proportion of spontaneous and indicated preterm births and assess risk factors for the subgroups of preterm birth during the period from 1991 to 2001. A population-based register study of all births occurring in Sweden from 1973 to 2001 registered in the Swedish Medical Birth Register was designed. The analysis of subgroups was restricted to the period 1991-2001. Gestational age was calculated using last menstrual period and best estimate. Odds ratio for preterm birth related to risk factors was calculated for the subgroups' spontaneous and indicated preterm birth. After an increase in the beginning of the 1980s, the preterm birth rate has decreased from 6.3% in 1984 to 5.6% in 2001 ( P < 0.0001). The proportion of multiple births born preterm of the total birth rate increased from 0.34% in 1973 to 0.71% in 2001 ( P < 0.0001). Spontaneous preterm births account for 55.2% and iatrogenic preterm births for 20.2% of all preterm births. The strongest association with maternal smoking in early pregnancy was found at gestational age <28 weeks and spontaneous preterm birth [odds ratio (OR) smoking versus no smoking: 1.55, 95% confidence intervals (CI): 1.42-1.69]. The strongest association for maternal age was found between gestational age <28 weeks and indicated preterm birth (OR 5-year increase: 1.34, 95% CI: 1.21-1.47). The preterm birth rate in Sweden has decreased since the mid 1980s. The composition of different subtypes of preterm birth in a Scandinavian low-risk population seems to be similar to populations with higher incidence of preterm birth and perinatal infections.