Study Background and Rationale: Given the large number of patients with a history of CVD, the substantial risk of further CVD morbidity and mortality, and the present and future availability of effective interventions on potentially key risk factors, it is important to characterize the burden of events (event rates and mortality) as well as the modifiable risk factors for subsequent CVD. The preponderance of existing data informing event rates in patients with existing CVD is derived from clinical trial populations. While such data is informative, clinical trials may underestimate event rates for a variety of reasons including: selection of patients with lower risk due to inclusion or exclusion criteria, disproportionate recruitment from high performing academic centers, or the higher quality of care received by clinical trial participants. Studies of risk factors for CV events have focused mostly on patients without existing CVD (the primary prevention setting) and thus prediction models focus on primary events as opposed to recurrent events. There may be important differences in predictors of primary events in an event-free population, and predictors of subsequent events in a population with existing CVD. Further, the relationship between these risk factors and outcomes may change over time from the primary CVD event, either due to changes in the natural history of disease, or to the use of various interventions. Such research requires population-based data with enough granularity in patient-level information (e.g. risk factors), a sufficient sample size, and a follow-up time long enough for such analyses to be feasible as well as potentially repeat measurements over time. Primary Objective(s): Objective 1: Estimate the rate of ACS (AMI, UA) or IS, plus CV and non-CV mortality, among patients with no history of CVD but who are at an elevated clinical risk due to a diagnosis of diabetes. Objective 2: Estimate the rate of ACS (AMI, UA), IS or HF, plus CV and non-CV mortality, among patients who have had ACS, IS or who have other clinically-evident CVD. Objective 3: Estimate the rate of CV and non-CV mortality among patients with HF secondary to an ACS event or other clinically-evident CVD Secondary Objective(s): Exploratory Objective 1: Evaluate the performance of existing risk equations to predict the rate of CVD-related events among patients with diabetes alone or with a history of ACS, IS or other clinically-evident CVD. Exploratory Objective 2: Estimate the association between LDL-C and risk of AMI, UA, and IS among patients with a history of CVD. Study Design/Type: Retrospective longitudinal descriptive cohort study Study Population or Data Resource: The study population will be derived from the SIDIAP database of electronic medical records in Catalonia and linked databases (e.g. hospital event data, and death records). Summary of Patient Eligibility Criteria: Patients must meet the following criteria: All patients: ? Age ? 30 years ? Continuous enrollment during a ? 2 year baseline period Cohort 1: ? Evidence of diabetes Cohort 2: ? Identified clinically-evident CVD Cohort 3: ? Incident non-fatal ACS or IS event Cohort 4: ? Incident HF subsequent to clinically-evident CVD or ACS or IS event. Outcome Variables: CVD events and CV and non-CV death. Follow-up: Data captured between July 2006 and December 2013 will be eligible for inclusion into this study. The baseline period and start of time at risk are defined differently for each study cohort. All cohorts must have continuous enrollment for ? 2yrs during which they do not experience an ACS or IS event. Sample Size: Minimun estimated sample size was calculated according to the number of expected cardiovascular events in the three cohorts of the study. Minimum sample sizes which will obtained from SIDIAP are the following: 47,000 (cohort 1), 22,500 (cohort 2), 26,000 (cohort 3), 10,000 (cohort 4).These estimated sample sizes will be obtained by far according to previous preliminary analysis in SIDIAP Statistical Considerations: Continuous and nominal variables will be described using number of observations, mean, standard deviation, median, interquartile range, and frequencies. For categorical variables, the number of observations, frequency and percent will be calculated. Incidence rates and 95% CIs will be calculated for ACS and IS and for CVD-related and non-CVD-related mortality The rate of HF will also be calculated for secondary prevention cohorts. Cause-specific hazard rates will be estimated by time-to-event models for each event individually. Other events, including death, will be treated as censored. Royston-Parmar models will be used, which are parametric survival models that use restricted cubic spline functions of time to permit more flexible hazard functions than are possible with Weibull, log-logistic, log-normal, and generalized gamma models.The adjusted hazard function will be based on a model including covariates.