Trends in cause-specific mortality among adults with and without diagnosed diabetes in the USA: an epidemiological analysis of linked national survey and vital statistics data

Summary

Background

Large reductions in diabetes complications have altered diabetes-related morbidity in the USA. It is unclear whether similar trends have occurred in causes of death.

Methods

Using data from the National Health Interview Survey Linked Mortality files from 1985 to 2015, we estimated age-specific death rates and proportional mortality from all causes, vascular causes, cancers, and non-vascular, non-cancer causes among US adults by diabetes status.

Findings

From 1988–94, to 2010–15, all-cause death rates declined by 20% every 10 years among US adults with diabetes (from 23·1 [95% CI 20·1–26·0] to 15·2 [14·6–15·8] per 1000 person-years), while death from vascular causes decreased 32% every 10 years (from 11·0 [9·2–12·2] to 5·2 [4·8–5·6] per 1000 person-years), deaths from cancers decreased 16% every 10 years (from 4·4 [3·2–5·5] to 3·0 [2·8–3·3] per 1000 person-years), and the rate of non-vascular, non-cancer deaths declined by 8% every 10 years (from 7·7 [6·3–9·2] to 7·1 [6·6–7·5]). Death rates also declined significantly among people without diagnosed diabetes for all four major mortality categories. However, the declines in death rates were significantly greater among people with diabetes for all-causes (p_interaction<0·0001), vascular causes (p_interaction=0·0214), and non-vascular, non-cancer causes (p_interation<0·0001), as differences in all-cause and vascular disease death between people with and without diabetes were reduced by about a half. Among people with diabetes, all-cause mortality rates declined most in men and adults aged 65–74 years of age, and there was no decline in death rates among adults aged 20–44 years. The different magnitude of changes in cause-specific mortality led to large changes in the proportional mortality. The proportion of total deaths among adults with diabetes from vascular causes declined from 47·8% (95% CI 38·9–58·8) in 1988–94 to 34·1% (31·4–37·1) in 2010–15; this decline was offset by large increases in the proportion of deaths from non-vascular, non-cancer causes, from 33·5% (26·7–42·1) to 46·5% (43·3–50·0). The proportion of deaths caused by cancer was relatively stable over time, ranging from 16% to 20%.

Interpretation

Declining rates of vascular disease mortality are leading to a diversification of forms of diabetes-related mortality with implications for clinical management, prevention, and disease monitoring.

Funding

None.

Introduction

Diabetes has been associated with an approximate 75% increase in mortality rate in adults, and the average 60 year old person diagnosed with diabetes loses 5 years of his or her life to the disease.1, 2, 3, 4 Although cardiovascular diseases have historically accounted for the largest number of those deaths, diabetes is associated with increased risk of death from a wide range of additional causes, including diabetes itself, renal disease, cancers, infections, liver disease, and external causes.1, 3, 4, 5, 6 Some population-based studies suggest that the character of diabetes-related complications could be changing.4, 7, 8 In US adults diagnosed with diabetes, the rate of cardiovascular disease-related complications declined more than other types of morbidities from 1990, to 2010, potentially increasing the relative importance of other complications and conditions, including chronic kidney disease, cancer, and ageing-related conditions.2, 8, 9, 10 It is possible that these shifts in morbidity are accompanied by changes in the causes of death among people with diabetes, as has been reported in the general population.¹¹ However, no studies have assessed cause-specific mortality among people with diabetes in the USA.⁷

National Vital Statistics data¹² suggest that the mortality rate attributed to diabetes decreased by 16% from 2000, to 2010. However, deaths caused by diabetes itself are difficult to interpret because of awareness, changing prevalence, and under-reporting of diabetes as a contributing cause.13, 14 Thus, determination of the causes and trends in death associated with diabetes requires follow-up mortality data based on cohorts of people with and without diabetes. In this study, we assembled data from two nationally representative datasets to assess the mortality burden associated with diagnosed diabetes, and to determine whether the magnitude and causes of death associated with diabetes are changing in concert with trends in diabetes complications.

Research in context

Evidence before this study

We searched PubMed for English-language reports of population-based studies of trends in causes of death in populations with diagnosed diabetes published from Jan 1, 1990, to June 1, 2017, using the terms “mortality trends”, and “diabetes mellitus” and separately “cause-specific mortality” and “diabetes mellitus”. Previous studies indicate that all-cause mortality among populations with diabetes has declined in several countries but no studies had examined trends in cause-specific mortality among people with diabetes or how those causes differ from populations without diagnosed diabetes.

Added value of the study

This study is the first nationally representative study of trends in the specific causes of death in adults with diagnosed diabetes and compares them to trends in those without diagnosed diabetes. Death rates due to vascular disease have declined more steeply among people with diabetes than in those without diabetes and decreased from about half of all deaths among people with diabetes in the 1990s, to about one-third in 2010–15. At the same time, the decrease in non-vascular, non-cancer deaths was modest and they now account for almost half of deaths in people with diabetes. Several of these causes, including renal disease, influenza and pneumonia, sepsis, and chronic liver disease, were significantly higher in people with diabetes than in those without.

Implications of all the available evidence

The disproportionate shift in causes of death from vascular and non-vascular causes has led to a diversification of causes of death, indicating a need to attend to clinical management, prevention, and monitoring of a more diverse range of diabetes-related conditions.