Objective: To compare trends, outcomes and risk factors for severe postpartum hemorrhage (PPH) due to uterine atony and retained tissue separately. Study design: This retrospective hospital-based study of severe PPH included deliveries from a 10-year cohort (2008–2017) and a four-year case-control group (2008–2011). Severe PPH was defined as an estimated blood loss of ≥1500 ml or a blood transfusion. Poisson regression was used to estimate the temporal trend in the 10-year cohort. Risk factors were investigated in the case-control group. We performed multinomial regression analysis to investigate associations between pregnancy characteristics and severe PPH caused by uterine atony and by retained tissue compared to controls without severe PPH. Results: During the 10-year study period, 2.7% of all deliveries were complicated by severe PPH. Uterine atony without concurring retained tissue was the cause in 55.4%, while retained tissue was listed as a cause in 32.2% of the cases. Among women who received ≥ four units of blood products, retained tissue caused 42.6% of cases, and severe PPH resulting in a hysterectomy was caused by retained tissue in 61.2% of cases. The rate of severe PPH caused by uterine atony significantly increased during the study period with an estimated annual percentage change of 8.6%, while the increase in severe PPH due to retained tissue was non-significant. Risk factors associated only with uterine atony were multiple pregnancy, macrosomia, Asian ethnicity and operative delivery, while induction of labor, augmentation of labor, use of anticoagulants and assisted reproduction were associated with both uterine atony and retained tissue. Conclusion: The observed increased rate of PPH in the study period was mainly driven by an increase in atonic PPH, while the rate of severe PPH caused by retained tissue remained stable. The proportion caused by retained tissue was highest among the most severe cases of PPH. The reason for the increase in severe PPH due to uterine atony, but not retained tissue was not clear, but we speculate that it may be a combination of increasing risk factors with increased awareness of PPH. [ABSTRACT FROM AUTHOR]