Trends in cardiovascular burden and mortality among adults with prediabetes in the US,1999

Abstract

Background

Trends in cardiovascular disease (CVD) risk factors and mortality have been reported among US adults with diabetes, but not among those with prediabetes. This study aimed to examine and compare the trends in CVD risk factors and mortality in US adults with diabetes and prediabetes.

Methods

In this serial, cross-sectional study, medical records of prediabetic patients from the National Health and Nutrition Examination Survey (NHANES) among adults aged 18 years or older were retrospectively reviewed.

Results

Data on 17,193 individuals including 7803 with diabetes and 9390 with prediabetes were analyzed from 1999–2000 to 2017–2018. A similar non-linear trend in the mean blood pressure (p = .991) and plasma fasting glucose (p = .068) was observed among the population with diabetes and prediabetes. The mean hemoglobin A1c decreased from 7.5% to 7.1% in diabetes and the trend was different from that in prediabetes (p = .004). Among both groups, a significant decline in the mean total cholesterol was identified while there was a difference in the trend (p = .003). The prevalence of hypertension remained largely the same for diabetes and it declined from 59.8% to 48.8% for prediabetes (p = .044). The prevalence of dyslipidemia increased from 40.4% to 53.5% in diabetes and it remained stable for prediabetes. The all-cause mortality decreased from 148.2 to 93.6 per 10,000 person-years between 1999–2006 and 2007–2014 in prediabetes and heart diseases mortality remained unchanged.

Conclusion

In the US from 1999 to 2018, the adults with prediabetes had different secular trends in the mean hemoglobin A1c and the prevalence of hypertension and dyslipidemia compared to those with diabetes. There was a significant reduction in all-cause mortality from 1999 to 2014 for the population with prediabetes. However, heart diseases mortality remained stable among them.

Keywords

Diabetes prediabetes cardiovascular disease

Introduction

The decrease in cardiovascular disease (CVD) mortality rates is a major public health success in the United States (US) over the last several decades.¹ This has been attributed to the control of CVD risk factors such as smoking and cholesterol levels and advances in healthcare management. However, CVD mortality remains to be the leading cause of death in the US.² Recent evidence also shows that the rate of CVD mortality flattened while the total number of death from CVD increased among the US general population in the last decade.^2,3 The reasons for this deceleration in the decline of CV mortality are not fully clarified.

cardiovascular disease closely relates to diabetes and common conditions coexisting with them (such as hypertension and dyslipidemia) are clear risk factors for CVD. Populations with diabetes have increased rates of cardiovascular burdens and mortality.^4,5 Numerous studies have shown the efficacy of aggressive management of CVD risk factors in preventing and slowing CVD in people with diabetes. Along with the improved quality of diabetes care, a significant reduction in all-cause and CVD mortality was observed among the US population with diagnosed diabetes between 1988 to 2015.^6,7

The prelude to diabetes is prediabetes and it can be described as a continuum from normoglycemia through worsening dysglycemia. It is estimated that nearly 34.5% of all US adults aged 18 years older had prediabetes in 2013–2016 in the US.⁸ Recent evidence suggested that adults with prediabetes have a higher prevalence of CVD burdens than those with normal glycaemic status.⁹ As a high-risk metabolic state for diabetes, prediabetes has also been proven to be associated with increased risk of all-cause and CVD mortality.¹⁰ However, no national studies of mortality trends in prediabetes are available, leaving the question of whether the control of CVD risk factors in this population has improved over the past decades.

To address these knowledge gaps, we analyzed the nationally representative data from the National Health and Nutrition Examination Surveys (NHANES) between 1999–2018 to examine and compare the trends in CVD risk factors and mortality in adults with diabetes and prediabetes over time.

Methods

Data source

The NHANES is a series of nationally representative, cross-sectional studies designed to be representative of the US civilian, noninstitutionalized population.¹¹ It was conducted by the National Center for Health Statistics of the Centers for Disease Control and Prevention. Participants are selected with the use of a complex, stratified, multistage probability-cluster sampling design for in-home interviews and visits to a mobile examination center. The design and methods of the NHANES have been described elsewhere.¹² The data from NHANES are publicly available to researchers around the world. We used data from the ten 2-years NHANES survey cycles conducted between 1999–2000 to 2017–2018. The National Center for Health Statistics approved the study protocols and written informed consent was obtained from each participant.

Data collection

This study aims to evaluate cardiovascular burdens. We selected the available variables from the NHANES which are related to cardiovascular burdens according to the literature.^4,9,13,14

During the in-home interview, a standardized questionnaire was used to collect demographic and health-related information.^11,12 Smoking status was self-reported based on responses to questionnaires regarding whether a participant was currently smoking or not. Medical history and medication were accessed during the interview. Participants were asked if they have ever been diagnosed with diabetes by a “doctor or other health professionals”. History of CVD including congestive heart failure, angina, myocardial infarction, or stroke was ascertained by self-report.

During the examination, weight, height, and waist circumference were measured. The body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared and the waist-to-height ratio was calculated waist circumference divided by height. Blood pressure (BP) was measured by trained staff after the participant rested quietly in a seated position for at least 5 min. Mean BP was calculated as the mean of three or four readings according to the standardized protocol. Blood samples were collected at the mobile examination center and tested at the central laboratories for the measurement of hemoglobin A1c and total cholesterol. Although different equipment was used for the measurement of hemoglobin A1c over time, calibration is unnecessary according to the recommendations from NHANES.¹⁵ Fasting plasma glucose (FPG) was only available in a subsample of survey participants after fasting for 8–24 h.

Diagnosed diabetes was defined as having a self-reported previous diabetes diagnosis; undiagnosed diabetes as a hemoglobin A1c value ≥6.5% or an FPG value ≥126 mg/dL among those without diagnosed diabetes; diabetes as either diagnosed or undiagnosed diabetes. Prediabetes was defined as a hemoglobin A1c value of 5.7–6.4% or an FPG value of 100–125 mg/dL among participants without diabetes and with both hemoglobin A1c and FPG values available.¹⁶ Hypertension was defined as the self-reported use of antihypertensive medications, systolic BP ≥130 mmHg or diastolic BP ≥80 mmHg; dyslipidemia as the self-reported use of lipid-lowering medications, or total cholesterol ≥240 mg/dL.

Ascertain of mortality

Mortality data of NHANES 1999–2014 were ascertained by linkage to the National Death Index through 31 December 2015. The ICD-10 was used to determine disease-specific death. According to the Codebook for the 2015 Public-Use Linked Mortality File, CVD mortality included mortality of heart diseases and cerebrovascular diseases.¹⁷ Given that mortality of cerebrovascular diseases was only available for NHANES 1999–2006, we investigate the trends in all-cause and heart diseases mortality in the current study.

Statistical analysis

Statistical analysis was conducted in R version 4.1.0 with the “Survey” package after accounting for the complex sampling design.¹⁸ To enable nationally representative estimation, we used the interview sampling weights for participants with diagnosed diabetes and the fasting sampling weights for adults without diagnosed diabetes so that the sum of sampling weights was added to the total US population.¹⁹ For each variable, participants with missing values were excluded from the analysis. All statistical tests were 2-sided, and p < 0.05 was considered statistically significant.

In the estimation of the prevalence of and trends in CVD risk factors among participants with diabetes and prediabetes, we pooled survey years into 4-years intervals (1999–2002, 2003–2006, 2007–2010, 2011–2014, and 2015–2018) to minimize the effects of small sample size and improve the precision of estimation. The age-standardized summary statistics were computed using the direct standardization methods to the 2000 US population with the following three categories: 18–44 years old, 45–64 years old, and 65 or more years. Means were estimated for continuous variables and prevalence for categorical variables. Linear or nonlinear secular trends were tested by linear regression (continuous variables) or logistic regression (dichotomized variables) with recoded pooled survey cycles (1–5) as a continuous variable. The nonlinearity of the trend was tested by adding a quadratic term or a logarithmic term into the regression model. We tested both the linear and nonlinear trend since the trend of some variables might be nonlinear. The p-value for the difference in secular trends among the population with diabetes and prediabetes was tested by adding an interaction term between survey cycles and diabetes status.

In the estimate of trends in mortality, we first pooled survey years into 8-years intervals due to the low all-cause and heart diseases mortality rate and we compared the difference in the mortality rate directly between 1999–2006 and 2007–2014. Considering that participants in 2011–2012 and 2013–2014 were only followed for a short period, we also compared the difference in mortality rate between 1999–2004 and 2005–2010. To minimize bias from differential follow-up, we limited follow-up to 10 years. Age-standardized mortality rates per 10,000 person-years were calculated. The number of person-years of follow-up time was calculated from the time of entry into the study until the date of death or the termination day of follow-up. Poisson regression was used to examine differences in rates across periods after adjusting for age, sex, and race. Furthermore, subgroup analyses by age (18–44 years old, 45–64 years old, and 65 or more years), sex (men and women), and race (Hispanic, non-Hispanic White, and non-Hispanic Black) were conducted.

Results

The analyses were restricted to non-pregnant adults aged 18 years or older. A total of 33,106 participants were identified from NHANES. Among them, 5113 were less than 18 years old and 696 were pregnant, resulting in 27,297 participants being non-pregnant adults aged 18 years or older. Among these participants, 7803, 9390, and 10,104 were identified as having diabetes, prediabetes, and normal glycemic status, respectively. The characteristics of participants in each survey was presented in Table 1. From 1999–2002 to 2015–2018, the mean age and sex distribution stayed largely stable among participants with diabetes, while a significant decrease in mean age and male proportion was observed among those with prediabetes (p for interaction, mean age: <0.05; the proportion of male: <0.01). The proportion of self-identified as non-Hispanic white decreased over time in both groups (p for interaction: 0.736). The mean BMI and waist-to-height ratio increased from 32.2 and 0.65 in 1999–2002 to 33.1 and 0.67 in 2015–2018 in individuals with diabetes. And for individuals with prediabetes, the mean BMI stayed stable and the mean waist-to-height ratio increased from 0.59 to 0.60 across the period. There were no differences in trends in the mean BMI and waist-to-height ratio among both groups (p for interaction, mean BMI: 0.205; mean waist-to-height ratio: 0.239). For both diabetes and prediabetes, a similar significant drop in the rate of current smoke was observed (p for interaction 0.829).

Table 1.

Cardiovascular disease risk factors burdens among US adults with diabetes and prediabetes, NHANES 1999–2018.

		1999–2002 (N = 4964)	2003–2006 (N = 4830)	2007–2010 (N = 6047)	2011–2014 (N = 5761)	2015–2018 (N = 5695)	Rate of change or odds ratio (confidence interval)	p for trend	p for interaction of the trend^a
No of diabetes (N = 7803)	No (proportion%) of missing cases	1159	1201	1766	1683	1994
Mean age, years	0 (0)	58.5 (0.9)	59.1 (0.7)	59.6 (0.5)	58.7 (0.6)	59.2 (0.6)	0.08 (−0.34–0.51)	0.699	0.013
Male, %	0 (0)	51.1 (2.1)	50.9 (1.9)	53.2 (1.7)	51.5 (1.9)	53.5 (2.1)	1.02 (0.97–1.08)	0.362	0.009
Non-hispanic white, %	0 (0)	64.8 (2.7)	66.2 (3.4)	63.0 (3.2)	58.8 (3.1)	59.1 (2.9)	0.91 (0.84–0.99)	0.033	0.736
Mean BMI, kg/m²	561 (7.2)	32.2 (0.5)	32.5 (0.4)	32.8 (0.3)	33.2 (0.3)	33.1 (0.4)	0.26 (0.03–0.49)	0.028	0.205
Mean waist-to-height ratio	974 (12.5)	0.65 (0.006)	0.65 (0.005)	0.66 (0.004)	0.67 (0.003)	0.67 (0.005)	0.005 (0.002–0.008)	0.002	0.239
Current smoking, %	28 (0.4)	19.6 (1.7)	19.6 (1.9)	16.4 (1.0)	15.5 (1.2)	15.3 (1.2)	0.92 (0.87–0.99)	0.019	0.829
Mean systolic pressure, mmHg	645 (8.3)	133.9 (1.0)	132.3 (0.9)	129.0 (0.9)	130.7 (0.9)	131.1 (0.6)	−5.3 (−8.23–−2.28)	<0.001	0.991
Mean hemoglobin A1c, %	626 (8.0)	7.5 (0.10)	7.0 (0.08)	7.0 (0.06)	7.2 (0.05)	7.1 (0.06)	−0.05 (−0.10–−0.00)	0.032	0.004
Plasma fasting glucose, mg/dL	3734 (47.9)	166.4 (4.5)	154.1 (3.1)	152.0 (2.2)	154.6 (2.4)	154.5 (2.1)	−12.3 (−23.4–−1.28)	0.043 ^b	0.068
Total cholesterol, mg/dL	750 (9.6)	208.5 (2.5)	198.7 (2.0)	187.6 (1.5)	184.1 (1.8)	183.1 (2.3)	−17.2 (−24.3–−10.1)	0.002 ^b	0.003
Hypertension, %	253 (3.2)	75.2 (2.1)	76.2 (1.7)	74.7 (2.5)	77.1 (1.6)	76.5 (2.0)	1.00 (0.93–1.08)	0.998	0.044
Dyslipidemia, %	434 (5.6)	40.4 (2.4)	50.4 (2.0)	51.7 (1.9)	53.5 (1.7)	53.5 (1.9)	1.50 (1.15–1.96)	0.025 ^b	0.010
History of CVD, %	104 (1.3)	24.1 (2.1)	28.0 (2.3)	24.4 (1.5)	24.3 (1.5)	24.3 (1.8)	0.97 (0.90–1.04)	0.402	0.991
No of prediabetes (9390)		1373	1415	2334	2016	2252
Mean age, years	0 (0)	51.1 (0.6)	50.2 (0.7)	48.8 (0.6)	49.6 (0.5)	48.8 (0.5)	−0.51 (−0.85–−0.17)	0.004
Male, %	0 (0)	59.6 (1.2)	57.9 (1.3)	56.7 (1.2)	54.6 (1.5)	54.3 (1.1)	0.94 (0.91–0.97)	<0.001
Non-hispanic white, %	9 (0)	72.1 (2.7)	71.5 (2.8)	69.1 (2.7)	65.9 (2.5)	63.3 (2.3)	0.90 (0.84–0.98)	0.010
Mean BMI, kg/m²	11 (1.3)	29.3 (0.2)	30.0 (0.2)	29.3 (0.2)	29.9 (0.3)	30.0 (0.2)	0.09 (−0.07–0.24)	0.260
Mean waist-to-height ratio	320 (3.4)	0.59 (0.003)	0.60 (0.003)	0.59 (0.003)	0.60 (0.003)	0.60 (0.004)	0.003 (0.001–0.005)	0.007
Current smoking, %	253 (2.7)	22.7 (1.4)	22.5 (1.9)	21.6 (1.1)	21.5 (1.4)	17.8 (1.3)	0.932 (0.884–0.983)	0.010
Mean systolic pressure, mmHg	341 (3.6)	127.6 (0.5)	126.6 (0.7)	122.3 (0.4)	123.2 (0.5)	124.3 (0.5)	−5.2 (−6.7–−3.6)	<0.001 ^b
Mean hemoglobin A1c, %	0 (0)	5.5 (0.02)	5.5 (0.02)	5.6 (0.02)	5.6 (0.01)	5.5 (0.01)	0.09 (0.04–0.13)	0.002 ^b
Plasma fasting glucose, mg/dL	0 (0)	105.0 (0.3)	105.2 (0.2)	105.0 (0.2)	104.5 (0.2)	106.4 (0.3)	−1.13 (−2.00–−0.26)	0.011 ^b
Total cholesterol, mg/dL	74 (0.8)	209.7 (1.7)	200.8 (1.1)	199.0 (1.0)	196.8 (1.1)	192.6 (1.4)	−3.6 (−4.6–−2.7)	<0.001
Hypertension, %	251 (2.7)	59.8 (1.6)	55.6 (2.0)	47.5 (1.4)	50.5 (1.7)	48.8 (1.7)	0.69 (0.54–0.88)	0.030 ^b
Dyslipidemia, %	60 (0.6)	32.1 (1.9)	30.0 (1.4)	30.4 (1.2)	32.2 (1.6)	29.6 (1.1)	1.00 (0.96–1.05)	0.917
History of CVD, %	390 (4.2)	9.8 (0.7)	10.1 (0.8)	8.6 (0.7)	9.5 (1.1)	7.9 (0.7)	0.97 (0.91–1.03)	0.323

All the values in the bracket (except for the No of missing cases and rate of change or odds ratio) represent standard error.

Bold font of the p values represents statistical significance.

BMI: body mass index; CVD: cardiovascular disease; NHANES: National health and nutrition examination survey.

^aTested by adding an interaction term between survey cycles and diabetes status into the linear or logistic mode.

^bTested for non-linear trend by adding a quadratic term into the linear or logistic model.

The mean systolic BP decreased from 133.9 mmHg in 1999–2002 to 129.0 mmHg in 2007–2010, then increased to 131.1 mmHg in 2015–2018 for participants with diabetes. A similar non-linear secular trend was observed for those with prediabetes (p for interaction: 0.991). For the population with diabetes, the mean hemoglobin A1c decreased from 7.5% to 7.1% between 1999–2002 and 2015–2018. A different trend in the mean hemoglobin A1c was observed for the population with prediabetes (p for interaction: <0.01). Similar non-linear changes in the mean PFG were observed in the population with diabetes and prediabetes (p for interaction: 0.068). Among both groups, a significant decline in the mean total cholesterol was identified while there was a difference in the trend (p for interaction: <0.01). The prevalence of hypertension remained largely the same from 1999–2002 to 2015–2018 for the population with diabetes and it declined from 59.8% to 48.8% during the same period for those with prediabetes (p for interaction: <0.01). Conversely, the prevalence of dyslipidemia increased from 40.4% in 1999–2000 to 53.5% in 2015–2018 for people with diabetes and it remained stable for those with prediabetes (p for interaction: <0.01). The reported history of CVD remained generally constant across time in both groups (p for interaction: 0.991).

For the participants from the survey of 1999–2000 to 2013–2014, the mean follow-up time was 8.3 years. The total number of death from all-cause was 2489 (1573 among participants with diabetes and 916 among participants with prediabetes) and the total number of death from heart diseases was 489 (334 among participants with diabetes and 155 from participants with prediabetes). Overall, the population with diabetes was at a higher risk of all-cause mortality compared to those with prediabetes from 1999–2006 to 2007–2014 (Table 2). This result remains broadly consistent in subgroups by sex, age, and race except for those aged 18–45 years in 2007–2014. For those with diabetes, there was a significant decrease in all-cause mortality in the overall participants (from 360.9 per 10,000 person-years in 1999–2006 to 273.6 per 10,000 person-years in 2007–2014) and different subgroups except for females. For those with prediabetes, a significant decrease in the all-cause mortality was observed in the overall participants (from 148.2 per 10,000 person-years in 1999–2006 to 93.6 per 10,000 person-years in 2007–2014) and in different subgroups except for those aged 18–44 years, Hispanic, and non-Hispanic black.

Table 2.

Difference in all-cause and heart diseases mortality among the US adults with diabetes and prediabetes between 1999–2006 and 2007–2014.

	All-cause mortality rate (95% CI)^a		Rate ratio (95% CI)	p value^b	Heart disease mortality rate (95% CI)^a		Rate ratio (95% CI)	p value^b
	1999–2006	2007–2014	Rate ratio (95% CI)	p value^b	1999–2006	2007–2014	Rate ratio (95% CI)	p value^b
All participants
Prediabetes	148.2 (132.4–164.0)	93.6 (77.4–109.8)	0.7 (0.5–0.8)	<0.001	20.2 (15.7–24.6)	18.9 (12.6–25.3)	1.0 (0.6–1.5)	0.914
Diabetes	360.9 (326.3–395.6)	273.6 (238.8–308.4)	0.7 (0.6–0.8)	<0.001	79.7 (62.5–96.8)	55.4 (40.7–70.1)	0.6 (0.4–0.9)	0.009
Rate ratio	2.4 (2.1–2.9)	2.9 (2.4–3.6)			4.0 (2.8–5.5)	2.9 (1.8–4.7)
Men
Prediabetes	139.7 (120.6–158.8)	90.2 (72.1–108.4)	0.7 (0.5–0.8)	0.001	20.6 (13.8–27.5)	22.9 (14.3–31.5)	1.1 (0.7–1.9)	0.642
Diabetes	408.8 (351.6–466.0)	275.8 (238.4–313.3)	0.6 (0.5–0.7)	<0.001	101.5 (72.2–130.9)	61.5 (41.5–81.4)	0.6 (0.4–0.9)	0.008
Rate ratio	2.9 (2.3–3.7)	3.1 (2.4–3.9)			4.9 (3.0–8.0)	2.7 (1.5–4.7)
Women
Prediabetes	160.2 (138.3–182.1)	97.9 (74.9–120.9)	0.7 (0.5–0.8)	0.001	19.5 (12.1–26.8)	13.9 (6.1–21.7)	0.8 (0.4–1.5)	0.416
Diabetes	313.9 (273.4–354.3)	271.2 (219.0–323.4)	0.8 (0.6–1.0)	0.054	58.1 (41.4–74.9)	48.8 (29.2–68.4)	0.8 (0.5–1.3)	0.294
Rate ratio	2.0 (1.6–2.4)	2.8 (2.1–3.7)			3.0 (1.9–4.8)	3.5 (1.6–7.9)
Age 18-44
Prediabetes	25.5 (13.6–37.4)	22.0 (9.2–34.8)	0.9 (0.4–1.8)	0.693	1.4 (0.0–3.1)	3.7 (0.0–8.8)	2.5 (0.4–16.0)	0.335
Diabetes	126.4 (73.9–179.0)	31.2 (10.0–52.4)	0.2 (0.1–0.6)	<0.001	27.2 (6.1–48.2)	6.6 (0.0–16.1)	0.2 (0.04–1.2)	0.083
Rate ratio	5.0 (2.7–9.1)	1.4 (0.6–3.6)			18.9 (4.6–78.0)	1.8 (0.2–13.5)
Age 45-64
Prediabetes	95.4 (75.3–115.4)	50.5 (33.1–68.0)	0.5 (0.4–0.8)	0.004	13.7 (7.4–19.9)	9.6 (0.8–18.4)	0.7 (0.3–2.1)	0.579
Diabetes	228.9 (190.1–267.6)	160.3 (120.2–200.4)	0.7 (0.5–0.9)	0.020	43.1 (22.1–64.5)	33.3 (13.8–54.8)	0.8 (0.4–1.7)	0.519
Rate ratio	2.4 (1.8–3.2)	3.2 (2.1–4.7)			3.2 (1.6–6.4)	3.6 (1.1–11.4)
Age over 65
Prediabetes	487.8 (436.1–539.5)	324.2 (252.3–396.0)	0.7 (0.5–0.9)	0.002	69.0 (50.0–87.9)	68.2 (43.4–93.1)	1.0 (0.6–1.6)	0.968
Diabetes	665.4 (611.0–719.8)	508.7 (433.1–584.3)	0.8 (0.6–0.9)	0.002	156.2 (121.6–190.7)	100.9 (73.0–128.7)	0.6 (0.5–0.9)	0.015
Rate ratio	1.4 (1.2–1.6)	1.6 (1.2–2.0)			2.3 (1.5–3.3)	1.5 (0.9–2.5)
Hispanic
Prediabetes	95.0 (53.8–136.2)	58.4 (37.5–79.2)	0.7 (0.4–1.1)	0.157	7.5 (2.2–12.8)	7.0 (0.6–13.4)	1.1 (0.3–3.8)	0.929
Diabetes	252.9 (198.3–307.4)	139.3 (96.9–181.8)	0.5 (0.3–0.6)	<0.001	71.5 (33.5–109.5)	17.6 (4.8–30.4)	0.2 (0.1–0.5)	<0.001
Rate ratio	2.7 (1.6–4.4)	2.4 (1.5–3.8)			9.5 (3.8–23.9)	2.5 (0.8–8.3)
Non-hispanic white
Prediabetes	169.6 (148.4–190.7)	108.4 (86.4–130.3)	0.6 (0.5–0.8)	<0.001	23.3 (17.6–29.0)	23.4 (15.2–31.6)	1.0 (0.7–1.6)	0.915
Diabetes	408.1 (359.5–456.8)	321.1 (272.9–369.3)	0.7 (0.6–0.9)	0.001	88.5 (64.0–113.1)	67.5 (47.1–87.9)	0.7 (0.5–1.0)	0.082
Rate ratio	2.4 (2.0–2.9)	3.0 (2.4–3.7)			3.8 (2.5–5.7)	2.9 (1.7–4.9)
Non-hispanic black
Prediabetes	117.6 (94.9–140.2)	73.4 (45.9–100.9)	0.7 (0.4–1.1)	0.100	22.6 (10.1–35.2)	15.7 (1.6–29.7)	0.8 (0.3–2.3)	0.641
Diabetes	308.1 (258.9–357.2)	257.5 (197.3–317.7)	0.7 (0.6–0.9)	0.013	70.6 (49.5–91.6)	53.1 (31.0–75.3)	0.6 (0.4–1.1)	0.109
Rate ratio	2.6 (2.0–3.4)	3.5 (2.3–5.3)			3.1 (1.6–6.0)	3.4 (1.2–9.3)

CI: confidence interval; US: United States.

^aMortality was expressed as events per 10000 person-years.

^bThe p value for the difference was calculated with Poisson regression after adjusting age, sex, and race.

Individuals with diabetes also had a higher risk of heart diseases mortality compared to those with prediabetes except for those aged 18–44 years, aged 65 or more years, and Hispanic in 2007–2014 Table 3. The heart disease mortality decreased from 79.7 per 10,000 person-years in 1999–2006 to 55.4 per 10,000 person-years in 2007–2014 in the population with diabetes, while this decrease was only significant in the subgroup of men, those aged 65 years or older, and Hispanic. Furthermore, for the population with prediabetes, there was no decrease in heart diseases mortality from 1999–2006 to 2007–2014 in the overall participants and different subgroups.

Table 3.

Comparisons between p values for trend with 4-years and 2-years interval.

	p for trend, diabetes		p for trend, prediabetes		p for interaction^a
	4-years	2-years	4-years	2-years	4-years	2-years
Mean age, years	0.699	0.567	0.004	0.002	0.013	0.006
Male, %	0.362	0.371	<0.001	<0.001	0.009	0.013
Non-hispanic white, %	0.033	0.036	0.010	0.007	0.736	0.568
Mean BMI, kg/m²	0.028	0.010	0.260	0.144	0.205	0.149
Mean waist-to-height ratio	0.002	0.001	0.007	0.004	0.239	0.184
Current smoking, %	0.019	0.012	0.010	0.004	0.829	0.808
Mean systolic pressure, mmHg	<0.001	<0.001	<0.001 ^b	<0.001 ^b	0.991	0.956
Mean hemoglobin A1c, %	0.032	0.023	0.002 ^b	0.001 ^b	0.004	0.002
Plasma fasting glucose, mg/dL	0.043 ^b	0.019 ^b	0.011 ^b	0.006 ^b	0.068	0.074
Total cholesterol, mg/dL	0.002 ^b	0.015 ^b	<0.001	<0.001	0.003	0.002
Hypertension, %	0.998	0.887	0.030 ^b	0.010 ^b	0.044	0.037
Dyslipidemia, %	0.025 ^b	0.008 ^b	0.917	0.805	0.010	0.018
History of CVD, %	0.402	0.349	0.323	0.269	0.991	0.963

^aTested by adding an interaction term between survey cycles and diabetes status into the linear or logistic mode.

^bTested for non-linear trend by adding a quadratic term into the linear or logistic model.

When limited the survey to 1999 and 2010, there was also a significant decline in all-cause and heart diseases mortality among participants with diabetes from 1999–2004 to 2005–2010. For those with prediabetes, the all-cause mortality declined while the heart diseases mortality remained largely stable across the same period (Figure 1).

Figure 1.

Difference in all-cause and heart diseases mortality among the US adults with diabetes and prediabetes between 1999–2004 and 2005–2010. RR: Rate ratio; CI: Confidence interval.

Discussion

The prevalence and absolute numbers of people with diabetes and prediabetes in the US grew substantially over the last two decades,⁸ making it an increasing health and economic burden. Reducing the CVD burden in diabetes is a major clinical imperative that should be prioritized since large benefits are witnessed when multiple CVD risk factors are addressed at the same time.^20,21 For prevention and management of CVD, risk factors including obesity, hypertension, dyslipidemia, and smoking should be systematically assessed in diabetes.²² The observations described in this nationally representative study of the US population from 1999–2000 to 2017–2018 offer a mix of welcome and troublesome findings. While a decrease in the rate of current smoke and mean total cholesterol was observed in the population with diabetes and prediabetes, the mean BMI and waist-to-height ratio increased and the prevalence of CVD remained stable among both groups. There were significant differences in the secular trends in the mean hemoglobin A1c and the prevalence of hypertension and dyslipidemia for the population with diabetes and prediabetes. Furthermore, consistent with previous studies,^23,24 it seems like the decline in the BP and glycemic among the population with diabetes and prediabetes plateaued or even increased after a decade of progress in risk factor control from 1999 to 2010. These findings of secular trends in cardiovascular burdens might be a public health concern.

Our study found that all-cause and heart diseases mortality continues to decrease after 2000 in individuals with diabetes. This result was consistent with the epidemiological analysis using data from the National Health Interview Survey Linked Mortality files in which all-cause death rates declined by 20% every 10 years among US adults with diabetes.⁷ A significant decline in the death rates from heart diseases was also identified in this study. Another similar study reported the trends in CVD mortality using the same data and the 10-years relative changes in mortality were significant for major CVD but not for the death of heart failure and arrhythmia.²⁵ However, unlike the current study in which the diabetes status was ascertained by both self-reported diagnosis and hemoglobin A1c and FPG values, these two studies relied on the self-reported diagnosis alone and the potential effects of undiagnosed diabetes on mortality were neglected.

To the best of our knowledge, the trends in mortality among participants with prediabetes have not been described in a nationally representative sample of the US. We found that between 1999–2006 and 2007–2014, there was a statistically significant decrease in all-cause mortality after adjusting for age, sex, and race. A similar trend was observed between 1999–2004 and 2005–2010. However, for heart diseases mortality, we didn’t find any evidence of decline across the periods. There are concerns about the current diagnostic thresholds for prediabetes which might be too low and labeling people as having prediabetes would lead to a medicalization of the condition.²⁶ However, a recent study has suggested that CVD burdens are highly prevalent in adults with prediabetes regardless of the different definitions used.⁹ Early identification of individuals with prediabetes might increase the opportunity for CVD burdens and thus reduce the rate of mortality.

There are sex differences in the trends of all-cause and heart diseases mortality among the population with diabetes. A statistically significant reduction in mortality was observed only in men. This result was consistent with the previous finding in which neither all-cause nor CVD mortality declined between 1971 to 1986 and 1988–2000. Factors including less aggressive medical management, differences in the pathophysiology of coronary heart disease, or differences in inflammatory and hormonal responses to CVD risk factors might lead to the sex difference in the mortality trend, and the lack of improvement among females remains to be a concern.⁶

The limitation of this study should be noted. First, this is a cross-sectional design. Even though successive surveys were used, they only reflected the profile of the population at the time of the survey. This design limits our ability to truly assess longitudinal associations between diabetes/prediabetes and excess CVD risk factors. Nevertheless, our study provided useful information about the prevalence and trends in CVD risk factors for US adults with diabetes and prediabetes. Second, the prediabetes in this study was defined by PFG and hemoglobin A1c, but not 2-h postprandial glucose concentrations since it’s challenging to collect these samples in a large epidemiological study. Another limitation was the relatively small samples in certain subgroups, which resulted in wide confidence intervals and therefore the point estimates should be interpreted with caution. Last, multiple statistical tests are performed in this study and the probability of finding at least one of them to be statistically significant just by chance might increase. However, we did not account for the problem of multiplicity during the statistical tests since there was currently no consensus on the need for multiplicity adjustment.^27,28

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Xingjian Zhang

References

Mensah

Wei

Sorlie

, et al. Decline in cardiovascular mortality: possible causes and implications. Circ Res 2017; 120: 366–380.

Virani

Alonso

Aparicio

, et al. Heart disease and stroke statistics—2021 update: a report from the American heart association. Circulation 2021; 143: e254–e743.

Shah

Lloyd-Jones

O’Flaherty

, et al. Trends in cardiometabolic mortality in the United States, 1999–2017. Jama 2019; 322: 780–782.

Saydah

Siegel

Imperatore

, et al. The cardiometabolic risk profile of young adults with diabetes in the US. Diabetes Care 2019; 42: 1895–1902.

Sarwar

Gao

Seshasai

SRK

, et al. Diabetes mellitus, fasting blood glucose concentration, and risk of vascular disease: a collaborative meta-analysis of 102 prospective studies. Lancet 2010; 375: 2215–2222.

Gregg

Cheng

, et al. Mortality trends in men and women with diabetes, 1971 to 2000. Ann Intern Med 2007; 147: 149–155.

Gregg

Cheng

Srinivasan

, et al. 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. Lancet 2018; 391: 2430–2440.

Centers for Disease Control and Prevention . National diabetes statistics report: estimates of diabetes and its burden in the United States. https://www.cdc.gov/diabetes/data/statistics-report/index.html (accessed 20 December 2021).

Ali

Bullard

Saydah

, et al. Cardiovascular and renal burdens of prediabetes in the USA: analysis of data from serial cross-sectional surveys, 1988–2014. Lancet Diabetes Endocrinol 2018; 6: 392–403.

10.

Cai

Zhang

, et al. Association between prediabetes and risk of all cause mortality and cardiovascular disease: updated meta-analysis. Bmj 2020; 370: m2297.

11.

Centers for Disease Control and Prevention . National health and nutrition examination survey. https://www.cdc.gov/nchs/nhanes/index.htm (accessed 23 December 2021).

12.

Statistics. NCfH . The National Health and Nutrition Examination Survey (NHANES) analytic and reporting guidelines. https://wwwn.cdc.gov/nchs/nhanes/analyticguidelines.aspx (accessed 28 May 2023).

13.

Yang

Cogswell

Flanders

, et al. Trends in cardiovascular health metrics and associations with all-cause and CVD mortality among US adults. Jama 2012; 307: 1273–1283.

14.

Patel

Taksler

, et al. Prevalence of elevated cardiovascular risks in young adults: a cross-sectional analysis of national health and nutrition examination surveys. Ann Intern Med 2017; 166: 876–882.

15.

Zipf

Chiappa

Porter

et al. National health and nutrition examination survey: plan and operations, 1999–2010. Vital Health Stat 1 2013; (56): 1–37.

16.

American Diabetes Association . 5. Prevention or delay of type 2 diabetes: standards of medical care in diabetes-2018. Diabetes Care 2018; 41: S51–S54.

17.

Centers for Disease Control and Prevention . Codebook for the 2015 public-use linked mortality file (LMF). https://www.cdc.gov/nchs/data/datalinkage/public-use-2015-linked-mortality-files-data-dictionary.pdf (accessed 23 December 2021).

18.

Survey: Analysis of Complex Survey Samples . https://cran.r-project.org/web/packages/survey/index.html (accessed 28 May 2023).

19.

Cheng

Kanaya

Araneta

MRG

, et al. Prevalence of diabetes by race and ethnicity in the United States, 2011–2016. Jama 2019; 322: 2389–2398.

20.

Nicolucci

De Berardis

Sacco

, et al. Primary prevention of cardiovascular diseases in people with diabetes mellitus: a scientific statement from the American heart association and the American diabetes association: response to Buse et al. Diabetes Care 2007; 30: e58–e570.

21.

Gæde

Lund-Andersen

Parving

H-H

, et al. Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med 2008; 358: 580–591.

22.

American Diabetes Association . 10. Cardiovascular disease and risk management: standards of medical care in diabetes-2021. Diabetes Care 2021; 44: S125–S150.

23.

Fang

Wang

Coresh

, et al. Trends in diabetes treatment and control in US adults, 1999–2018. N Engl J Med Overseas Ed 2021; 384: 2219–2228.

24.

Zhu

Bundy

, et al. Trends in cardiovascular risk factors in US adults by race and ethnicity and socioeconomic status, 1999–2018. JAMA 2021; 326: 1286–1298.

25.

Cheng

Imperatore

Geiss

, et al. Trends and disparities in cardiovascular mortality among US adults with and without self-reported diabetes, 1988–2015. Diabetes Care 2018; 41: 2306–2315.

26.

Yudkin

Montori

. The epidemic of pre-diabetes: the medicine and the politics. Bmj 2014; 349: g4485.

27.

Taljaard

Van den Heuvel

, et al. An introduction to multiplicity issues in clinical trials: the what, why, when and how. Int J Epidemiol 2017; 46: 746–755.

28.

Streiner

Norman

. Correction for multiple testing: is there a resolution? Chest 2011; 140: 16–18.

Trends in cardiovascular burden and mortality among adults with prediabetes in the US,1999–2018

Abstract

Background

Methods

Results

Conclusion

Keywords

Introduction

Methods

Data source

Data collection

Ascertain of mortality

Statistical analysis

Results

Discussion

Footnotes

Declaration of conflicting interests

Funding

ORCID iD

References