Age at type 2 diabetes diagnosis and the risk of mortality among US population (2024)

Type 2 diabetes (T2DM) is a prevalent public health problem in the world today, characterized by persistent hyperglycemia due to insulin resistance and destruction of insulin-producing β-cells¹. Epidemiological surveys in 2021 estimated that the number of people with T2DM worldwide was approximately 537million, and this figure is projected to increase to 783million by 2045². As T2DM progresses, it leads to microcirculatory and macrovascular complications (e.g., cardiovascular disease, chronic kidney disease) as well as non-complicated critical illnesses, contributing to a high mortality rate of approximately 6.7million deaths worldwide in 2021².

T2DM accounts for approximately 90% of the total number of diabetic patients, and historically it was predominantly seen in middle-aged and older adults³. However, in recent decades, due to the growth of obesity rates, individual factors, and changes in lifestyle behaviors, it is not uncommon to find a younger population with T2DM⁴. The 2017 SEARCH data reported 0.67 (95% CI 0.63,0.70) cases of T2DM per 1,000 among adolescents in the U.S⁵. T2DM diagnosed in individuals under the age of 40 years is also referred to as “early onset adult T2DM”. Early-onset adult T2DM worldwide has been increasing year by year and is estimated to account for more than 15% of all adult T2DM patients worldwide⁶. Early-onset T2DM is considered more aggressive, with poorer outcomes and a higher relative risk of diabetes-related complications and death^7,8. This is attributed to a bidirectional impact relationship between type 2 diabetes and cardiovascular disease, leading to vascular damage and inflammatory responses that further increase the risk of cardiovascular disease⁹. Therefore, it is worthwhile to investigate the relationship between age at diagnosis of T2DM mellitus and cardiovascular disease and all-cause mortality.

Previous studies of the age at onset of T2DM among Americans have explored age-related trends, or the association between the age at onset and the risk of ensuing cardiovascular disease^10,11. Our investigation on the association between age at diagnosis of T2DM and cardiovascular disease and all-cause mortality in a U.S. population is, to our knowledge, the first of its kind. In this study, we examined the relationship between age at diagnosis of T2DM and cardiovascular disease and all-cause mortality using data gathered from The National Health and Nutrition Examination Survey (NHANES) database, which is representative of the entire U.S. population, with the expectation that it will provide evidence for standardized diagnosis, therapy, and intervention strategies for T2DM.

Study population and design

NHANES is conducted by the National Center for Health Statistics, a division of the Centers for Disease Control and Prevention (CDC), to evaluate the health and nutritional status of the non-institutionalized population in the United States. NHANES implements a complex survey design and utilizes population-specific sample weights to yield nationally representative data on the non-institutionalized civilian population biennially. The study protocol for NHANES experiences review and approval by the Ethics Review Board of the National Center for Health Statistics Research, and written informed consent is obtained from all participants. For more information about NHANES, visit the NHANES website at https://www.cdc.gov/nchs/nhanes/index.htm.

In the present study, data was garnered from ten successive NHANES cycles, ranging from 1999 to 2018, covering a comprehensive number of 101,316 subjects. The exclusion criteria implemented are as follows: (1) participants below the age of 18, (2) pregnant participants, (3) participants with missing essential diabetes diagnostic data, (4) participants without diabetes, (5) participants identified with type 1 diabetes, and (6) participants who missed data of age at T2DM diagnosis or mortality. Diabetes was defined as meeting any of the following criteria: (1) self-reported history of diabetes, (2) taking diabetic medication to lower blood sugar, (3) fasting plasma glucose level ≥ 126mg/dL, or (4) a hemoglobin A1c level ≥ 6.5%. Type 1 diabetes was defined by self-report of diagnosis before age 30 years and on insulin monotherapy. The screening regimen is graphically represented in Fig.1.

Age at T2DM diagnosis

Age at T2DM diagnosis was determined based on self-reported history. For participants with previously undiagnosed diabetes, we considered their age at the interview as the age at diagnosis. In this study, age at T2DM diagnosis was categorized as < 40 years, 40–59 years, and ≥ 60 years.

Ascertainment of mortality

To ascertain the mortality status in the follow-up population, we utilized the NHANES public-use linked mortality file updated up to December 31, 2019. This file was connected with the National Death Index (NDI) employing a probability matching algorithm by the National Center for Health Statistics (NCHS). Additionally, disease-specific deaths were identified using the International Statistical Classification of Diseases, 10th Revision (ICD-10), with heart diseases (codes 054–068) classified by the NCHS for our study.

Covariables

Sociodemographic characteristics were collected using standardized household questionnaires. The details collected included age, sex, race, marital status, education level, family poverty income ratio (PIR), smoking status, and alcohol consumption. Race was classified into Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, and other races. Marital status was divided into two groups: unmarried (including those who were never married, divorced, separated, or widowed) and married (including those who were married or living with a partner). Education level was divided into three categories: less than high school, high school or equivalent, or college and above. PIR was grouped into four categories: < 1.3, 1.3–3, 3–5, and ≥ 5. Smoking status was categorized into three groups: never smoker (having smoked fewer than 100 cigarettes in a lifetime), former smoker (having smoked at least 100 cigarettes in a lifetime but not currently smoking), and current smoker (having smoked at least 100 cigarettes in a lifetime and continuing smoking). Participants consuming a minimum of 12 alcoholic drinks per year were considered to have alcohol consumption. Body mass index (BMI) was calculated as weight in kilograms divided by height squared in square meters (kg/m²). Estimated glomerular filtration rate (eGFR) was ascertained using the Chronic Kidney Disease Epidemiology Collaboration creatinine equation. Laboratory data, including glycohemoglobin, plasma fasting glucose, uric acid, and creatinine were collected based on established protocols. Hypertension was defined using the following items: (1) systolic blood pressure ≥ 140mm Hg, (2) diastolic blood pressure ≥ 90mm Hg, (3) taking anti-hypertensive medication, or (4) self-reported history of hypertension. The duration of diabetes was determined by calculating the interval between the age at the interview and the age at diagnosis. Hyperuricemia was characterized by a uric acid level of ≥ 7mg/dL. The use of insulin or anti-diabetic drug was according to the questionnaire by asking participants, “Are you now taking insulin? /Are you now taking diabetic pills to lower your blood sugar?“. The diagnosis of cardiovascular diseases (CVD) was established using the Monetary Choice Questionnaire by asking participants, “Has someone ever told you that you had coronary heart disease/angina pectoris/heart attack/stroke?“.

Statistical analysis

In accordance with the guidelines provided by the National Center for Health Statistics, all estimates were computed while taking into account the NHANES sample weights. Continuous variables are displayed as mean (95%CI), while categorical variables are expressed as counts and percentages. For comparing continuous variables among various groups, a survey-weighted linear regression model was employed, and for categorical variables, a survey-weighted chi-square test was used. Multivariate cox regression analysis with sample weights was conducted to assess the association of age at T2DM diagnosis with cardiovascular and all-cause mortality. Three adjusted regression models were established: Model 1 was adjusted for sex and race. Model 2 was adjusted for Model 1 plus education level, Marital status, PIR. Model 3 was adjusted for Model 2 plus smoking status, alcohol consumption, BMI, eGFR, hypertension, hyperuricemia, duration of diabetes, and CVD. Furthermore, to calculate the standardized mortality rates, we used propensity score matching to match each diagnostic age group (< 30, 30–39, 40–49, 50–59, 60–69, and ≥ 70) with a corresponding non-diabetic control group at a 1:1 ratio, based on the subjects’ current actual age, gender, and race.

To investigate potential exposure-effect relationships between the age at T2DM diagnosis and mortality, restricted cubic spline regression after full adjustments was utilized. In subgroup analyses, possible modifications of the association between the age at T2DM diagnosis and mortality were assessed across various variables, including sex (male vs. female), race (Mexican American vs. other Hispanic vs. Non-Hispanic white vs. Non-Hispanic black vs. other races), BMI (< 25 vs. 25-29.9 vs. ≥ 30), eGFR (< 60 vs. 60-89.9 vs. ≥ 90), smoking status (never vs. former vs. current), alcohol consumption (yes vs. no), hypertension (yes vs. no), and CVD (yes vs. no). To ensure the robustness of data analysis, sensitivity analyses were conducted in participants with previously diagnosed diabetes.

All statistical analyses were carried out using R, version 4.3.1 (R Foundation). Significance was determined using a two-tailed p-value, with values less than 0.05 considered statistically significant.

Baseline characteristics of study population

In the current study, a total of 8654 participants were included in the final analysis. Table1 presents the baseline characteristics of the study subjects, stratified by quartiles of age at T2DM diagnosis. Generally, the mean age was 59.61 years (95%CI: 59.16 to 60.05). Furthermore, the mean age at diagnosis was 51.71 years (95%CI: 51.22 to 52.19) (Table S1). Non-Hispanic whites constituted the majority of participants, comprising 35.19% of the sample, with males representing 51.59%. The median age of participants diagnosed with diabetes at age < 40, between 40 and 59, and > 60 are 44.04, 57.59, and 72.24, respectively. The follow-up periods for these participants are 67,554 person-years, 67,609 person-years, and 67,625 person-years, respectively. Subjects diagnosed with T2DM at a younger age exhibited several distinct characteristics. They were more likely to be females and younger. Additionally, they had a higher prevalence of current smokers, a higher percentage of Mexican American and Non-Hispanic black, a longer duration of diabetes, and a higher rate of insulin usage. Economically, these subjects were often characterized by lower income status and educational attainment. Furthermore, this group had higher levels of glycohemoglobin, plasma fasting glucose, BMI, eGFR. Moreover, a comparatively lower prevalence of diverse complications such as hypertension, hyperuricemia, CVD (all p < 0.05), was observed among participants diagnosed with T2DM at a younger age.

During 67,625 person-years of follow-up, a total of 2,582 all-cause deaths and 722 CVD-related deaths were recorded. In the preliminary unadjusted model, an older age at T2DM diagnosis was associated with a higher all-cause mortality (for each 1-unit increment, HR 1.04; 95% CI, 1.03–1.04) and CVD mortality (for each 1-unit increment, HR 1.04; 95% CI, 1.03–1.04). However, following comprehensive adjustments, the original positive correlation between age at T2DM diagnosis and mortality shifted. It was observed that older age at T2DM diagnosis was associated with lower all-cause mortality (for each 1-unit increment, HR 0.98; 95% CI, 0.97–0.99) and CVD mortality (for each 1-unit increment, HR 0.98; 95% CI, 0.97–1.00). In other words, as the age at T2DM diagnosis decreases, the corresponding risk of both all-cause and CVD mortality escalates. When the age at diagnosis of T2DM was assessed as a categorical variable, younger age categories were found to be associated with an increased risk of all-cause (< 40 vs. ≥60, HR, 2.72; 95% CI, 1.83–4.05) and CVD mortality (< 40 vs. ≥60, HR, 2.74; 95% CI, 1.31–5.74) after full adjustments as shown in Table2. Notably, a linear relationship between the age at T2DM diagnosis and all-cause mortality was detected (p for nonlinearity = 0.103) (p for nonlinearity = 0.235) (Fig.2A). However, cardiovascular disease mortality reaches a turning point at age 54 years, showing a decreasing trend until age 54, followed by a gradual upward trend (Fig.2B). Similarly, for patients with a self-reported history of T2DM, this turning point advances to age 51 years, further corroborating the heightened risk of death amongst T2DM patients (Fig.2D).

Additionally, to validate the effectiveness of our method and the accuracy of our results, we have matched a group of general population controls based on age, gender, and ethnicity, and calculated the standardized mortality rates. Table3 displays the all-cause and cardiovascular mortality rates for each diabetes diagnosis age group and their corresponding matched control (MC) groups; it also presents the incidence rate ratios (IRR) between the diabetes group and the MC group. As shown in Table3, when the age at diabetes diagnosis is less than 60 years, the IRR between the diabetes group and the MC group decreases with the increasing age at diabetes diagnosis. However, when the age at diabetes diagnosis is greater than 60 years, there is no statistically significant change in the IRR between the diabetes group and the MC group with the increasing age at diabetes diagnosis.

Sensitivity analyses

Considering the possible influence of defining the age at T2DM diagnosis, we conducted a sensitivity analysis centered on individuals previously diagnosed with diabetes. The analysis underscored that younger age at T2DM diagnosis is associated with both all-cause mortality (< 40 vs. ≥60, HR, 2.59; 95% CI, 1.73–3.87) and CVD mortality (< 40 vs. ≥60, HR, 2.23; 95% CI, 1.04–4.80) (Table4). Additionally, the restricted cubic splines highlighted a linear relationship between age at T2DM diagnosis and all-cause mortality in this demographic group (p for nonlinearity = 0.235) (Fig.2C).

Stratified analyses

Details of subgroup analyses are presented in Figs.3 and 4. These analyses did not reveal significant interactions in subgroups based on gender, race, BMI, eGFR, alcohol consumption, and CVD (p > 0.05). However, a significant interaction was observed in the smoking subgroup (current vs. never, p for interaction < 0.05) and hypertension subgroup (yes vs. no, p for interaction < 0.05). The relationship between the age at T2DM diagnosis and all-cause and CVD mortality was notably stronger in the current smokers and hypertensive population.

Observations from historical trends in epidemiological studies of T2DM exhibit a gradual decrease in the age of diagnosis, a pattern that aligns with many disease trends escalated by societal factors¹². Before, T2DM is considered a disease of the rich in China, and with economic development, the incidence of diabetes elevates, while the age of onset may gradually become younger. However, a recent study suggests that the age of onset of T2DM in Americans has stabilized in the 21st century, which may be related to the gradual stabilization of the food industry and lifestyle habits¹¹. Previous studies have probed the link between age at diagnosis of diabetes and the risk of all-cause mortality, macrovascular disease, and microvascular disease, and have shown an inverse relationship (p < 0.001)¹³. In this study, we investigated the association between age at T2DM diagnosis and cardiovascular disease and all-cause mortality in the US population. Similar to previous findings in other countries, we found that a younger age of T2DM diagnosis corresponds to a heightened risk of cardiovascular disease and all-cause mortality^{14,15,16,17,18,19}. Our findings bolster this understanding further, with the robustness of results affirmed by meticulous sensitivity analyses. We further performed subgroup analyses and found that the association between age at diagnosis of T2DM and all-cause and cardiovascular disease mortality was significantly stronger in those who smoked or had hypertension, a finding that may provide additional clinically relevant guidance for our conclusions.

The potential mechanisms linked to the elevated risk of cardiovascular disease and all-cause mortality with early onset T2DM warrant exploration. Therefore, it may be necessary for clinical practice to implement strategies aimed at early diagnosis and treatment, and to adapt screening and treatment approaches to prevent complications and death in these patients²⁰. As reflected in Table1, there is a higher proportion of obesity (BMI > 30) among individuals diagnosed with T2DM at a younger age. Also, the percentage of current smoker was high. This suggests that younger diabetics may be more obese and more inclined to smoke than non-diabetics. In other words, the impact of obesity and smoking is notably more significant with early onset T2DM, which in turn leads to a relative increase in the risk of cardiovascular and other diseases. In contrast, the high rates of comorbid hyperlipidemia, hypertension, and nephropathy in patients with diabetes at older ages are consistent with this reality. Furthermore, individuals diagnosed with T2DM at early ages appear to have higher smoking rates and lower socioeconomic status, which are also independently critical risk factors for cardiovascular disease. We also discovered a smaller proportion (65.71%) of hypoglycemic drug use among those below 40 in the T2DM population, a trend that may be associated with economic factors and compliance, and thus requires deeper discussion.

From the results of exposure-effect analysis, it can be inferred that both in the entire study population and in patients diagnosed with T2DM as previously reported, the risk of all-cause mortality progressively declines with the age at T2DM diagnosis, exhibiting a linear correlation. However, a pivotal point in cardiovascular mortality was observed at age 54, illustrated by a gradual yearly decrease until that age, followed by a slow rising trend after reaching age 54. For the T2DM patients based on self-reported history, this turning point occurs earlier, at age 51, further corroborating the increased death risk associated with T2DM. The rise in cardiovascular mortality post-turning point may be attributable to an amplified risk of cardiovascular death, potentially due to the increased prevalence of accompanying cardiovascular diseases with age.

We identified 2,582 all-cause deaths and 722 cardiovascular disease deaths during 67,625 person-years of follow-up. Disease prevention strategies that aim at delaying the onset of T2DM may be of great benefit in terms of the relationship between the age at diagnosis of T2DM and the risk of death. Moreover, adolescents will be a key target to protect. Given that the earlier the onset of the disease, the worse the prognosis, it is noteworthy that there is a lack of research on young T2DM patients, whilst such patients are not particularly rare. Based on SEARCH 2014–2015 data, the incidence of T2DM in children and adolescents in the United States is not low, with 13.8 cases per 100,000 adolescents/year, and the incidence rate is increasing every year²¹. Similarly, the prevalence of type 3 diabetes in adolescents in other countries and regions of the world is not optimistic²². Firstly, maintaining a healthy diet, lifestyle, and a healthy BMI may be a strategy to guard against the early onset of T2DM. Secondly, it is imperative for the scientific community to enhance its understanding of the pathophysiology of T2DM in adolescents, which will help to identify at-risk populations for preventive measures. Furthermore, there is a need to conduct further research to explore onset at different ages of diagnosis and to inform the design of age-specific interventions.

These results carry significant clinical ramifications for epidemiological studies and T2DM prevention protocols, underlining the necessity of T2DM population screening and more intensive risk modification in individuals diagnosed at a younger age or newly diagnosed with T2DM. Although pre-existing guidelines offer direction on risk factor management in patients diagnosed with T2DM early in life, these guidelines necessitate further consolidation in a constructive and uniform way, and application to disease prevention and treatment remains a looming challenge^23,24.

Our study expanded meaningfully on relevant published data by: (1) assessing a wide spectrum of all-cause mortality and cardiovascular disease death outcomes throughout a specified follow-up period; and (2) employing sensitive analyses and subgroup analyses to augment the persuasiveness of our findings. (3) This is the first study to examine the association between age at diagnosis and cardiovascular disease and all-cause mortality in a U.S. population. Consequently, the outcomes of this investigation deepen our comprehension of early-onset T2DM, furnish evidence-based backing for the screening and prevention of T2DM, and delineate potential future management approaches.

At the same time, it is crucial to acknowledge that there are limitations to this study. Firstly, as the study is based on a U.S. population, caution must be exercised when applying its findings to other geographies due to the specifics of the data representation. Secondly, most of the data were collected from questionnaires, which may have some recall bias. Thirdly, since the type of diabetes was not recorded in NHANES, we excluded type 1 diabetes based on previously reported methods. However, it is important to recognize that this method may be not perfect and could have resulted in the presence of type 1 diabetics in the sample.

Author notes

1. These authors contributed equally: Hong-Jin Zhang, Jie Feng and Xiang-Tao Zhang.

Authors and Affiliations

1. Department of Cardiovascular Medicine, First Affiliated Hospital of Gannan Medical University, Ganzhou, 341000, Jiangxi, China

   Hong-Jin Zhang&Hong-Zhou Zhang

2. Department of Cardiology, The Second Affiliated Hospital, Jiangxi Medical College, Nanchang University, Nanchang, 330006, China

   Jie Feng&Xiang-Tao Zhang

Contributions

HZ.Z was responsible for the entire project and revised the draft. HJ.Z, J.F, and XT.Z performed the data extraction and statistical analysis, interpreted the data, and drafted the first version of the manuscript. All authors contributed to the article and approved the submitted version.

Corresponding author

Correspondence to Hong-Zhou Zhang.

Competing interests

The authors declare no competing interests.

Ethics statement

This study was reviewed and approved by the National Center for Health Statistics Research Ethics Review Board, and written informed consent was obtained from all NHANES participants.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions