Little is known about differences in patient characteristics before and after implementation of the new definition of sepsis (Sepsis-3) and whether the new definition is affecting clinical practice in intensive care units. To examine and compare the clinicoepidemiologic characteristics of patients with sepsis or septic shock before and after implementation of Sepsis-3. In this population-based cohort study, a nationwide registration database in South Korea was used to identify patients with sepsis or septic shock. Patients admitted to hospitals from 2012 to 2015 constituted the Sepsis-2 group, and patients admitted from 2017 to 2020 constituted the Sepsis-3 group. The study involved 443 217 patients, of whom 170 660 (38.5%) were in the Sepsis-2 group and 272 557 (61.5%) were in the Sepsis-3 group. The mean (SD) age was 73.3 (14.5) years in the Sepsis-2 group and 75.5 (14.5) years in the Sepsis-3 group. The intensive care unit admission rate was higher in the Sepsis-2 group (34.6%, 59 081 of 170 660) than in the Sepsis-3 group (21.3%, 57 997 of 272 557). Multivariable Cox regression analysis showed that 1-year all-cause mortality was 21% lower in the Sepsis-3 group than in the Sepsis-2 group (hazard ratio, 0.79; 95% CI, 0.78-0.79; P < .001). Implementation of the Sepsis-3 definition was associated with an increased number of patients with sepsis. Other findings suggested that patients in the Sepsis-2 group had more severe illness, with increased 1-year all-cause mortality, compared with those in the Sepsis-3 group.