DM and PM are associated with substantial morbidity and mortality. We aimed to examine recent trends.Using The Health Improvement Network, we identified patients with incident DM/PM (defined by ≥1 Read diagnosis code) aged 18-89 years with ≥1 year of continuous enrolment prior to the cohort entry date and up to 10 comparators matched on age, sex and entry year. The cohort was divided in two based on the year of DM/PM diagnosis: the early cohort (1999-2006) and late cohort (2007-2014). We calculated multivariable hazard ratios (HR) for death using a Cox-proportional hazards model and multivariable rate differences (RD) using an additive hazard model.We identified 410 DM cases (mean age: 58 years, 66% female) and 407 PM cases (mean age: 59 years, 61% female). Both DM cohorts had excess mortality compared with the comparison cohorts (71.5 vs 12.9 deaths/1000 person-years [PY] in the early cohort and 49.1 vs 10.4 deaths/1000 PY in the late cohort). The multivariable HRs were 7.51 (95% CI: 4.20, 13.42) in the early cohort and 5.42 (95% CI: 3.11, 9.45) in the late cohort (P-value for interaction = 0.59), and multivariable RDs were 56.2 (95% CI: 31.8, 81.2) in the early cohort and 36.3 (95% CI: 19.6, 53.0) in the late cohort (P-value for interaction = 0.15). A similar trend existed in PM.The premature mortality gap in DM/PM has not considerably improved in recent years, highlighting an unmet need for therapeutic improvement.