Life years lost associated with diabetes: An individually matched cohort study using the U.S. National Health Interview Survey data

Zhiqiang Wang, Meina Liu

Previous estimates of life-years lost to diabetes are highly inconsistent. This study provided the updated estimates of life-years lost to diabetes in the United States. Each of a nationally representative sample of 21,829 adults with diabetes in the U.S. National Health Interview Survey 1997-2009 was individually matched to one without diabetes by age, sex, race, survey year, BMI, smoking status, pre-existing cardiovascular disease and pre-existing cancer. All-cause mortality from original surveys to 31 December 2011 and median survival ages were estimated for those with diabetes and their matched controls. Overall median survival age for adults with diabetes was 10.5years shorter than that for matched controls without diabetes. Estimated life-years lost associated with diabetes decreased with increasing age at diagnosis from 20.0years for those diagnosed before age 20years to no difference for those diagnosed after 80years. Hazard ratios for mortality decreased from 3.03 (95% CI: 2.41, 3.80) for those with diabetes diagnosed before 20years to 1.04 (95% CI: 0.78, 1.39) for those diagnosed after 80years. The estimate of life-years lost associated with diabetes was much higher among those with pre-existing cardiovascular disease (20.3years) than among those without cardiovascular disease (8.5years). The effect of diabetes on survival depends on age at first diagnosis of diabetes and the presence of pre-existing diseases. The life-years lost are higher for those with diabetes diagnosed at younger ages. This study provided the updated estimates of life-years lost associated with diabetes in the United States.