Introduction
Cardiovascular disease (CVD) encompasses a group of disorders that impact the heart and blood vessels, resulting in disorders such as coronary artery disease, heart failure, and stroke. There were 153.2 million disability-adjusted life years and 6.9 million diet-related CVD deaths worldwide in 2019, representing increases of 34.3% and 43.8%, respectively, since 1990 [1].
CVD is a result of several risk factors, including poor dietary habits, a sedentary lifestyle, smoking, excessive alcohol consumption, and hereditary predisposition [2–4]. Hypertension [5], diabetes [6], and obesity [7] are significant and important contributors. Early detection and management of risk factors along with medical intervention, when necessary, are crucial in mitigating the impact of CVD on individuals and public health.
The prevalence of type 2 diabetes (diabetes for short) has been steadily increasing on a global scale, presenting a significant public health challenge [8]. The International Diabetes Federation (IDF) projects that 537 million people between the ages of 20 and 79 years will develop diabetes by 2021. Current trends predict that the number of people with diabetes will rise to 643 million by 2030 and 783 million by 2045 [9]. Diabetes significantly increases the risk of developing cardiovascular disease (CVD), indicating a close and complex link between the 2 conditions. The impact of diabetes on the cardiovascular system stems from a combination of factors associated with the metabolic disorder [6, 10–12]. While those who have diabetes are more likely than those without it to get CVD [10, 11], the associations between diabetes and the incidence of CVD in different blood pressure, glycated haemoglobin A1c (HbA1c), and lipid levels are not well-established. We conducted a prospective cohort study in Guiyang, China, to gain a better understanding of how blood pressure, HbA1c, and lipid levels contribute to the development of CVD in diabetes patients. Our study aimed to examine the associations between diabetes and CVD risk based on blood pressure, HbA1c, and lipid status.
Material and methods
Study participants
The study participants were recruited from the Guiyang subcentre of the China Cardiometabolic Disease and Cancer Cohort (4C) Study, a multicentre, population-based, prospective cohort study [12, 13]. Between 2011 and 2012, local resident registration systems in Guiyang recruited 10,140 individuals aged 40 years or older. After excluding nonparticipants and those with cardiovascular diseases at baseline, 9672 participants remained. All participants received an in-person visit during the follow-up survey, extending from 2014 to 2016. However, only 7236 participants (74.8%) attended the follow-up visit. At baseline, all study participants had complete data on covariates, with none reporting a history of CVD. We determined whether they experienced a cardiovascular event during the follow-up visit. As depicted in Figure 1, the final analysis comprised 7197 participants. This study was approved by the Shanghai Jiaotong University Medical Ethics Committee at Ruijin Hospital. Each participant in the trial provided written informed consent.
Data collection
We collected data at local community clinics in the mornings. Trained interviewers used standardised questionnaires to gather information regarding the participants’ medical history, demographic characteristics, lifestyle factors (including cigarette smoking and alcohol consumption), and family history. We classified the education levels of the participants as follows: junior school or below and high school or further education. Participants were classified according to their marital status as either married or unmarried, including those who were separated, solitary, widowed, or divorced.
We classified each person’s smoking status as never, former, or current. Participants who consumed alcohol at least once a week in the past 6 months were considered current consumers of alcohol.
Trained nurses assessed the height and body weight of the participants in accordance with a standardised protocol. We computed the body mass index (BMI) by dividing the weight in kilograms by the height in square metres.
The participants’ systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured using an automated electronic device (OMRON Model HEM-752 FUZZY, Dalian, China) while they were seated and had remained still for a minimum of 5 minutes, following a 30-minute abstinence from alcohol, coffee, tea, smoking, and exercise. The average of 3 measurements was utilised in the analysis.
After an overnight fast of a minimum of 10 hours, every participant underwent a 2-hour oral glucose tolerance test (OGTT) using 75 grams of glucose. We obtained blood samples at the 0- and 2-hour time points of the experiment. We measured fasting and 2-hour plasma glucose concentrations within 2 hours of blood sample collection using a glucose oxidase or hexokinase method; all procedures adhered to the rigorous quality control regimen. The researchers obtained capillary whole blood samples from the participants’ fingertips using the Hemoglobin Capillary Collection System (Bio-Rad Laboratories, Hercules, CA, USA). We then transported and preserved the samples at temperatures ranging from 2°C to 8°C. Four weeks after the collection, we determined the levels of HbA1c using the Variant II Hemoglobin Testing System (Bio-Rad Laboratories) in conjunction with high-performance liquid chromatography at the central laboratory of the Shanghai Institute of Endocrine and Metabolic Disease.
The central laboratory measured the serum levels of low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides, and total cholesterol using an auto-analyser (ARCHITECT ci16200, Abbott Laboratories, Chicago, IL, USA).
Assessment of diabetes
Based on the American Diabetes Association 2010 criteria [16], participants were defined as diabetic at the baseline as per the following criteria: fasting plasma glucose level ≥ 7.0 mmol/L (126 mg/dL), OGTT 2-h plasma glucose level ≥ 11.1 mmol/L (200 mg/dL), HbA1c concentration ≥ 6.5%, or a self-reported prior diagnosis of diabetes by medical professionals [16]. The other participants did not have diabetes.
Definition of ideal blood pressure, HbA1c, and lipid level
While the American Heart Association (AHA) has proposed 7 ideal cardiovascular health metrics to promote optimal cardiovascular health in the population [14], our study focused on examining the impact of blood pressure, HbA1c, and lipid levels on CVD. We estimated each participant’s glucose metabolism using HbA1c instead of fasting plasma glucose. This modification aligns with the strict HbA1c targets outlined in the most recent clinical practice guidelines from the American Diabetes Association [15]. Based on previous research, our study prioritised the optimal management of the following 3 factors: (1) SBP < 130 mm Hg and DBP < 80 mm Hg [16, 17]; (2) HbA1c < 6.5% [12, 15]; and (3) triglycerides < 1.70 mmol/L, total cholesterol < 5.20 mmol/L, HDL-cholesterol > 1.04 mmol/L, and LDL-cholesterol < 3.37 mmol/L [21–23].
Identification of cardiovascular events
Our study’s primary outcome was the composite of incidental fatal or nonfatal CVD events. This composite was composed of outcomes related to heart failure treated in hospitals, stroke, myocardial infarction, and cardiovascular mortality. We have already provided comprehensive definitions of various CVD events [12]. Detailed information on deaths and clinical outcomes was collected from the local vital registers of the National Disease Surveillance Point System and the National Health Insurance System. To ensure precision and uniformity, 2 outcome adjudication committee members independently verified every clinical event and assigned potential reasons for death. Discrepancies were resolved through discussions with fellow committee members. Crucially, the outcome adjudication committee members remained oblivious to each participant’s baseline clinical characteristics, thereby enhancing the objectivity and impartiality of the outcome verification process.
Statistical analysis
We summarised the baseline characteristics of participants, classified by diabetes status, using means with standard deviations (SDs) for continuous variables and numbers with percentages for categorical variables. We censored the data pertaining to participants who died, received a CVD diagnosis, or reached the end of their follow-up, whichever occurred first. Each participant’s person-time from the date of enrolment to the date of censorship was computed. The differences in baseline characteristics among diabetes status groups were evaluated using a t-test and a chi-squared (X2) test.
During the follow-up period, we calculated the disparities in subsequent CVD between participants with and without diabetes using the log-rank test and Kaplan–Meier method. The relationship between diabetes and later CVD events was assessed using the multivariable Cox proportional hazards model. We adjusted Model 1 for sex, age, education, and marital status. We further adjusted Model 2 for alcohol drinking, smoking status, BMI, family history of cancer and diabetes, hypertension, hyperlipidaemia, cancer, liver diseases, pancreatitis, gastrointestinal diseases, gallbladder diseases, lung diseases, and renal diseases. The hazard ratios (HRs) and 95% CI were computed for each model to assess the association between diabetes and CVD events. Subgroups were formed based on blood pressure, HbA1c, and lipid levels, and the relationship between diabetes and CVD was examined. We also investigated the association of ideal blood pressure, ideal HbA1c, and ideal lipid levels with CVD events, stratified by diabetes diagnosis.
A 2-sided P value below 0.05 was considered statistically significant. We performed the statistical analyses using the IBM SPSS software, version 25.0 (New York, United States).
Results
Diabetes was diagnosed in 590 (8.20%) of the 7197 participants. The median (25th–75th percentiles) baseline age was 57 (52–63) years, and the median (25th–75th percentiles) follow-up duration was 3.15 (3.07–3.21) years. The baseline characteristics of the study participants are categorised in Table 1 according to whether or not they were diagnosed with diabetes at the commencement of the study. Participants with diabetes at baseline exhibited a higher average age, a greater proportion of males, a larger percentage of individuals with overweight or obesity, a higher likelihood of exposure to both former and present smoking, a higher percentage of having a family history of cancer and diabetes, and a greater proportion of chronic diseases (including hypertension, hyperlipidaemia, liver diseases, gallbladder diseases, gastrointestinal diseases, and renal diseases) compared to the participants without diabetes. We discovered that the participants with diabetes were less likely to have optimal blood pressure, HbA1c level, and lipid levels (Tab. 2).
|
Characteristic |
Diagnosed with diabetes |
p-value |
|
|
Yes (n = 590) |
No (n = 6607) |
||
|
Age, years |
60.89 ± 7.49 |
57.46 ± 7.73 |
< 0.01 |
|
Males, n (%) |
237 (40.2) |
1673 (25.3) |
< 0.01 |
|
Education, n (%) |
0.43 |
||
|
Junior school or below |
327 (55.4) |
3550 (53.7) |
|
|
High school or above |
263 (44.6) |
3057 (46.3) |
|
|
Marital status, n (%) |
0.62 |
||
|
Married |
499 (84.6) |
5,638 (85.3) |
|
|
Not married* |
91 (15.4) |
969 (14.7) |
|
|
BMI |
0.02 |
||
|
< 23.0 kg/m2, n (%) |
204 (34.6) |
2670 (40.4) |
|
|
23.0–24.9 kg/m2, n (%) |
158 (26.8) |
1658 (25.1) |
|
|
≥ 25 kg/m2, n (%) |
228 (38.6) |
2279 (34.5) |
|
|
Smoking status |
< 0.01 |
||
|
Nonsmoker, n (%) |
406 (68.8) |
5211 (78.9) |
|
|
Former smoker, n (%) |
63 (10.7) |
385 (5.8) |
|
|
Current smoker, n (%) |
121 (20.5) |
1011 (15.3) |
|
|
Alcohol drinking |
0.01 |
||
|
Non/former drinker, n (%) |
445 (75.4) |
4656 (70.5) |
|
|
Current drinker, n (%) |
145 (24.6) |
1951 (29.5) |
|
|
Family history of cancer and diabetes, n (%) |
227 (38.5) |
1719 (26.0) |
< 0.01 |
|
Hypertension, n (%) |
176 (29.8) |
1075 (16.3) |
< 0.01 |
|
Hyperlipidaemia, n (%) |
135 (22.9) |
599 (9.1) |
< 0.01 |
|
Cancer, n (%) |
25 (4.2) |
198 (3.0) |
0.10 |
|
Liver diseases, n (%) |
70 (11.9) |
364 (5.5) |
< 0.01 |
|
Gallbladder diseases, n (%) |
181 (30.7) |
1486 (22.5) |
< 0.01 |
|
Pancreatitis, n (%) |
12 (2.0) |
92 (1.4) |
0.21 |
|
Gastrointestinal diseases, n (%) |
38 (6.4) |
672 (10.2) |
0.04 |
|
Renal diseases, n (%) |
46 (7.8) |
313 (4.7) |
0.01 |
|
Lung diseases, n (%) |
33 (5.6) |
286 (4.3) |
0.15 |
|
Index |
Diagnosed with diabetes |
p-value |
|
|
Yes (n = 590) |
No (n = 6607) |
||
|
Blood pressure |
|||
|
SBP [mmHg] |
127.85 ± 18.32 |
120.73 ± 18.19 |
< 0.01 |
|
DBP [mmHg] |
77.15 ± 10.43 |
77.03 ± 10.89 |
0.79 |
|
Optimal blood pressure, n (%) |
293 (49.7) |
3843 (58.2) |
< 0.01 |
|
Glucose profile |
|||
|
Fasting glucose [mmol/L] |
8.87 ± 2.96 |
5.93 ± 1.10 |
< 0.01 |
|
OGTT-2 h glucose [mmol/L] |
13.85 ± 4.61 |
8.14 ± 3.04 |
< 0.01 |
|
HbA1c (%) |
7.69 ± 1.65 |
6.08 ± 0.74 |
< 0.01 |
|
Optimal HbA1c level, n (%) |
112 (19.0) |
5609 (84.9) |
< 0.01 |
|
Lipid profile |
|||
|
LDL cholesterol [mmol/L] |
2.54 ± 0.86 |
2.64 ± 0.88 |
0.01 |
|
HDL cholesterol [mmol/L] |
1.15 ± 0.34 |
1.25 ± 0.37 |
< 0.01 |
|
Triglycerides [mmol/L] |
2.00 ± 1.52 |
1.72 ± 1.30 |
< 0.01 |
|
Total cholesterol [mmol/L] |
4.52 ± 1.26 |
4.61 ± 1.24 |
0.09 |
|
Optimal lipid level, n (%) |
144 (24.4) |
2130 (32.2) |
< 0.01 |
A total of 134 CVD incidents were reported during 22,587 person-years of follow-up. Ten participants had multiple CVD events; 109 suffered strokes; 21 experienced heart failure; and 14 developed myocardial infarction. As shown in Figure 2, the participants with diabetes exhibited a higher cumulative hazard risk of CVD events during the follow-up period compared to those without diabetes (p < 0.01). After taking into account the potential confounding factors (Supplementary File — Table S1), the multivariable adjusted HR for CVD events linked to diabetes in all the participants was 2.37 (95% confidence interval (CI): 1.38–4.08). Compared to participants without diabetes, diabetes had different risk effects on CVD events: the hazard ratio was 2.24 (95% CI: 1.39–4.21) for people with hypertension, 1.23 (95% CI: 0.54–2.82) for people with normal blood pressure, 2.06 (95% CI: 1.01–4.19) for people with high HbA1c, 2.91 (95% CI: 1.31–6.44) for people with normal HbA1c, 2.19 (95% CI: 1.29–3.70) for people with hyperlipidaemia, and 1.26 (95% CI: 0.47–3.41) for people with normal lipid levels (Tab. 3).
|
Category |
N |
Person-years |
Cases |
HR (95% CI) |
|
|
Model 1 |
Model 2 |
||||
|
Total |
|||||
|
No diabetes |
6,607 |
20,746 |
109 |
1.00 (Ref.) |
1.00 (Ref.) |
|
Diabetes |
590 |
1,841 |
25 |
2.04 (1.31–3.17) |
1.80 (1.16–2.81) |
|
High blood pressure |
|||||
|
No diabetes |
2,761 |
8,668 |
55 |
1.00 (Ref.) |
1.00 (Ref.) |
|
Diabetes |
297 |
923 |
18 |
2.73 (1.59–4.66) |
2.42 (1.39–4.21) |
|
Optimal blood pressure |
|||||
|
No diabetes |
3,846 |
12,078 |
54 |
1.00 (Ref.) |
1.00 (Ref.) |
|
Diabetes |
293 |
918 |
7 |
1.24 (0.57–2.75) |
1.23 (0.54–2.82) |
|
High HbA1c level |
|||||
|
No diabetes |
998 |
3,144 |
16 |
1.00 (Ref.) |
1.00 (Ref.) |
|
Diabetes |
478 |
1,495 |
18 |
2.16 (1.09–4.28) |
2.06 (1.01–4.19) |
|
Optimal HbA1c level |
|||||
|
No diabetes |
5,609 |
17,602 |
93 |
1.00 (Ref.) |
1.00 (Ref.) |
|
Diabetes |
112 |
346 |
7 |
3.4 (1.57–7.37) |
2.91 (1.31–6.44) |
|
High lipid level |
|||||
|
No diabetes |
4,477 |
14,059 |
72 |
1.00 (Ref.) |
1.00 (Ref.) |
|
Diabetes |
446 |
1,389 |
20 |
2.32 (1.40–3.83) |
2.19 (1.29–3.70) |
|
Optimal lipid level |
|||||
|
No diabetes |
2,130 |
6,687 |
37 |
1.00 (Ref.) |
1.00 (Ref.) |
|
Diabetes |
144 |
453 |
5 |
1.46 (0.57–3.75) |
1.26 (0.47–3.41) |
Furthermore, we explored the associations between ideal blood pressure, HbA1c, and lipid level with CVD events among all participants, including those with and without diabetes. We found that optimal blood pressure maintenance significantly reduces the incidence of CVD events, especially in individuals with diabetes (HR: 0.35, 95% CI: 0.14–0.86). The effects of optimal blood pressure were not statistically significant in any of the participants, including those without diabetes (Supplementary File — Table S2). Contrary to expectations, ideal HbA1c and lipid levels did not exhibit a reduced risk of CVD events among all participants, those without diabetes, and those with diabetes (Supplementary File — Tables S3 and S4).
Discussion
Through this prospective cohort study involving 7197 adults from the Guiyang subcentre, we have demonstrated once again that those with diabetes are more likely to experience CVD events. Researchers discovered that adults with diabetes were more likely to experience CVD events under conditions of elevated blood pressure and lipid levels [18]. However, adults with diabetes did not have a statistically significant increase in the risk of CVD events when their blood pressure and lipid levels were optimal. Regardless of their HbA1c levels, adults with diabetes face an increased risk of CVD. Among patients with diabetes, only those with optimal blood pressure were associated with a lower risk of CVD compared to those with hypertension. Our results emphasise the significance of optimising the metabolic profile of cardiovascular risk factors to lower the incidence of CVD as an integral component of effective diabetes management [19].
Our findings consistently highlight that participants with diabetes face a notably higher risk of CVD events. This observation aligns with existing evidence pointing to diabetes as a substantial risk factor for CVD development [20–22]. The complex interplay between diabetes and cardiovascular health involves various metabolic, inflammatory, and vascular factors [23–25]. Understanding and addressing these intricate mechanisms is crucial for developing effective preventive strategies. The implications of our findings underscore the importance of targeted interventions and heightened vigilance in managing cardiovascular health among individuals with diabetes. It prompts further exploration into customised preventive strategies, highlighting the necessity of all-encompassing care plans that consider the unique challenges that diabetes presents in relation to CVD risk.
Our findings also support a significant finding that optimal management of the metabolic profile of cardiovascular risk factors, particularly lipid and blood pressure levels, could significantly counteract the adverse effects of diabetes on incident CVD events. Previous studies have repeatedly demonstrated that, in adults with diabetes, optimisation of cardiovascular risk variables can result in substantial reductions in severe CVD events [12, 30–33]. Our findings provide similar valuable insights into the potential mitigation of diabetes’s adverse effects on incident CVD events. The observed correlation underscores the importance of personalised strategies aimed at achieving and maintaining optimal blood pressure and lipid levels among individuals with diabetes [26–31]. Diabetes remained, unexpectedly, independently associated with the CVD risk even at optimal HbA1c levels. This finding may be largely attributed to the clustering of other cardiovascular risk factors associated with hyperglycaemia [40]. Our team’s previous study [12] supports this finding. These findings suggest that, in addition to conventional diabetes therapy, prioritising metabolic control could be essential in lowering the general burden of CVD in this group. While our observations are compelling, the precise mechanisms responsible for these associations remain unclear. Future research into the intricate underlying mechanisms may facilitate the development of tailored and effective interventions for preventing diabetes-related CVD events.
The prospective study design, the well-established definitions of exposure and CVD outcomes, and the thorough analyses involving people with or without optimal blood pressure, HbA1c, and cholesterol levels were among this study’s strong points. However, the study has several limitations. First, we only conducted a single follow-up in our cohort for analysis. The comparatively short follow-up time might limit the statistical power to identify CVD events. Second, this investigation could not explain temporal variations in the parameters because it only assessed baseline blood pressure, glucose, and lipid profiles. Third, after controlling for confounding factors such as BMI and smoking, we evaluated the cardiovascular effects of diabetes using only 3 optimal cardiovascular health metrics. Furthermore, the influence of unmeasured and residual variables, including psychological state, drugs, and genetic predisposition, could not be entirely ruled out.
Conclusions
In conclusion, our findings indicate that diabetes significantly increases the risk of developing CVD events. Crucially, this risk can be mitigated through the optimal management of the metabolic profile of cardiovascular risk factors. Our research emphasises the necessity of early intervention to improve metabolic profiles in patients with diabetes to prevent CVD.
Acknowledgements
The authors thank all the study participants and research staff involved in the Guiyang subcentre of the China Cardiometabolic Disease and Cancer Cohort Study, as well as the members of the endocrinology department at Ruijin Hospital, for their support. We would like to thank Dr. Guoyong Ding from Shandong First Medical University for her earnest assistance in revising and editing this manuscript.
Ethics statement
The Ethical Review Committee of Ruijin Hospital (RUIJIN-2011-14) reviewed and approved the studies involving human participants. The participants provided their written informed consent to participate in this study.
Funding
This study was supported by the China Cardiometabolic Disease and Cancer Cohort Study.
Conflicts of interest
The authors declare no conflicts of interest to disclose.
Author contributions
Y.Z. conceived and designed the study. Y.Z. and N.P. drafted the manuscript. Y.Z., D.W., and M.Z. conducted the analyses. N.P. contributed to the statistical analysis. Q.Z. and L.S. contributed to the funding acquisition. Q.Z. and L.S. contributed to supervision, had full access to all the study’s data, and took responsibility for the integrity and accuracy of the data analysis. All authors approved the final version of the manuscript.
