Predictors of mortality over 8 years in type 2 diabetic patients: Translating Research Into Action for Diabetes (TRIAD)

OBJECTIVE

To examine demographic, socioeconomic, and biological risk factors for all-cause, cardiovascular, and noncardiovascular mortality in patients with type 2 diabetes over 8 years and to construct mortality prediction equations.

RESEARCH DESIGN AND METHODS

Beginning in 2000, survey and medical record information was obtained from 8,334 participants in Translating Research Into Action for Diabetes (TRIAD), a multicenter prospective observational study of diabetes care in managed care. The National Death Index was searched annually to obtain data on deaths over an 8-year follow-up period (2000–2007). Predictors examined included age, sex, race, education, income, smoking, age at diagnosis of diabetes, duration and treatment of diabetes, BMI, complications, comorbidities, and medication use.

RESULTS

There were 1,616 (19%) deaths over the 8-year period. In the most parsimonious equation, the predictors of all-cause mortality included older age, male sex, white race, lower income, smoking, insulin treatment, nephropathy, history of dyslipidemia, higher LDL cholesterol, angina/myocardial infarction/other coronary disease/coronary angioplasty/bypass, congestive heart failure, aspirin, β-blocker, and diuretic use, and higher Charlson Index.

CONCLUSIONS

Risk of death can be predicted in people with type 2 diabetes using simple demographic, socioeconomic, and biological risk factors with fair reliability. Such prediction equations are essential for computer simulation models of diabetes progression and may, with further validation, be useful for patient management.In 2007, diabetes was the seventh leading cause of death in the U.S. with >71,000 death certificates listing diabetes as the underlying cause of death (1,2). Diabetes is reported as the underlying cause of death on 10–15% of death certificates of decedents with diabetes and is reported anywhere on the death certificate for 40% of decedents with diabetes (3–5). Since a diagnosis of diabetes cannot be ascertained from death certificates, the best way to examine mortality in people with diabetes is to use a prospective observational cohort design and follow diabetic participants until death. Unfortunately, few such studies have been undertaken in nationally representative populations of people with diabetes. Although many investigators have assessed risk factors for cardiovascular mortality, there are few published prediction models. Such models are needed to construct computer models to simulate the progression of diabetes, and may, with further validation, be useful for clinical decision making.Translating Research Into Action for Diabetes (TRIAD), which began in 1999, was a prospective observational study of people with diabetes enrolled in managed care health plans in the U.S. The primary objective of TRIAD was to study how health plan, provider, and patient characteristics impact the processes and outcomes of diabetes care. TRIAD enrolled ～12,000 diabetic patients from eight geographically distinct regions in the U.S. We previously reported on the demographic, socioeconomic, and biological risk factors for 4-year mortality in the TRIAD population (6). Since our study was published, two studies have reported prediction equations for mortality in patients with type 2 diabetes. One was performed in Chinese patients living in Hong Kong (7). The other was limited to patients initially prescribed a single oral hypoglycemic agent (8).The objective of this study was to assess longer-term mortality in a large, diverse, U.S. population with type 2 diabetes and to create prediction equations for all-cause, cardiovascular, and noncardiovascular mortality among people with clinically diagnosed type 2 diabetes.