Association between changes in systolic blood pressure and the incidence of diabetes mellitus: a retrospective study based on the Korean National Health Screening Cohort
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
As the global prevalence of diabetes mellitus continues to increase, it is important to identify its risk factors and implement preventive approaches. This study aimed to investigate the association between changes in systolic blood pressure (SBP) and the incidence of diabetes.
Methods
Data from 152,547 participants, who underwent two consecutive health checkups between 2002 and 2003, and 2004 and 2005, and included in the Korean National Health Insurance Service Health Screening Database, were reviewed. Participants were divided into three groups according to change in SBP: decrease (≥10 mm Hg); no change (<10 mm Hg); and increase (≥10 mm Hg). Cox proportional hazard regression models for diabetes incidence were constructed to evaluate adjusted hazard ratio (HR) with corresponding 95% confidence interval (CI).
Results
The median follow-up was 14.3 years, and 26,352 patients with diabetes were identified. Compared to those with no change in SBP, the adjusted HRs for decrease and increase among males were 1.06 (95% CI, 1.02–1.10) and 1.10 (95% CI, 1.06–1.14), and 1.06 (95% CI, 1.00–1.12) and 1.08 (95% CI, 1.02–1.13) for females, respectively. After stratifying data according to SBP at baseline, the HRs for decrease in males and females were as follows: normotensive group, 1.16 (95% CI, 1.08–1.24) and 1.11 (95% CI, 1.02–1.21); and prehypertensive group, 1.14 (95% CI, 1.09–1.20) and 1.19 (95% CI, 1.10–1.29), respectively.
Conclusion
Changes in SBP were associated with a risk for diabetes.
According to the International Diabetes Federation, the number of individuals with diabetes globally was estimated to reach 589 million by 2024, corresponding to 11.1% of adults 20 to 79 years of age, and projected to increase to 853 million by 2050 [1]. With this increase, health expenditures due to diabetes amount to US$1.015 trillion, an increase of 338% over the past 17 years [1]. Therefore, it is necessary to identify risk factors for diabetes and to implement preventive approaches to modifiable risk factors.
Modifiable risk factors for diabetes include obesity, smoking, alcohol consumption, physical inactivity, prolonged sedentary time, unhealthy diet, low high-density lipoprotein cholesterol levels, high triglyceride levels, and high systolic blood pressure (SBP) [2,3]. Several studies have reported an association between increased SBP and the occurrence of diabetes [4,5]. Thus, SBP control can be considered to be helpful in preventing diabetes and cardiovascular complications. However, other studies investigating the relationship between SBP and diabetes have reported inconsistent results. Some have reported a significant association between changes in SBP, either increases or decreases through control, and the risk for diabetes [6,7], whereas others have found no such association [8]. These mixed results suggest that changes in SBP may affect the risk for diabetes even if baseline SBP is not high. Furthermore, the effects of antihypertensive medications have not been completely controlled [4-8].
As such, the present study aimed to identify how changes in SBP may be related to the incidence of diabetes using data from a nationally representative database, including prescription history.
Methods
Data source and study population
The present study used information from the Korean National Health Insurance Service Health Screening (NHIS-HEALS) Cohort Study database, which comprises 10% of national health insurance subscribers 40 to 79 years of age from 2002 to 2003, among other national health screening examinees. The database includes information from 514,866 individuals, including demographic characteristics, personal and family medical history, lifestyle behaviors, laboratory results from health checkups and medical institute usage, disease information, such as diagnostic codes, and prescriptions from claims data were included [9]. The Korean NHIS provides free health checkups for individuals ≥40 years of age every alternate year.
A flow-diagram illustrating the participant selection process, based on predefined inclusion and exclusion criteria, is presented in Figure 1. First, 334,937 participants who underwent their first health checkup between 2002 and 2003, and their second between 2004 and 2005, were identified, among whom 182,390 were excluded, leaving 152,547 who were ultimately included in the present study.
Ethics
This study was approved by the Institutional Review Board of the Chungbuk National University Hospital (CBNUH; Cheongju, South Korea, 2023-03-018) and adhered to the guidelines of the Declaration of Helsinki (1975). Due to the retrospective design of the study and the use of anonymized participant data, requirements for informed consent were waived by the Ethics Committee.
Operational definition and study duration
The primary outcome was the occurrence of diabetes, which was defined as the fulfillment of any of the following criteria: fasting blood glucose level ≥126 mg/dL; ≥1 diagnosis of diabetes (International Classification of Diseases, 10th Revision [i.e., “ICD-10”] code E10–14); and prescription history of antidiabetic drugs (metformin, sulfonylurea, thiazolidinedione, inhibitor of dipeptidyl peptidase 4, sodium-glucose cotransporter two inhibitor, insulin, glucagon-like peptide one receptor agonist, and/or their combination[s]) for >90 days. Blood pressure (BP) data were based on information from two consecutive health checkups. BP was measured at medical institutions designated as providers of the National Health Screening Program, in accordance with the methods specified in the “Health Screening Implementation Guidelines” (Notice No. 2025-5) issued by the Mistry of Health and Welfare. To compare the risk for diabetes with changes in BP, study participants were divided into three groups according to the difference in SBP between the first and second examination dates: decrease (SBP decrease ≥10 mm Hg); no change, change <10 mm Hg; and increase, SBP increase ≥10 mm Hg.
The start of the study was the second examination date, and the end date was the date of the occurrence of diabetes or death. If neither occurred, the end date was the last health checkup, hospital visit, or prescription.
Confounding variables
Body mass index (BMI) was calculated as body weight (kg) divided by height squared (m2). Information regarding lifestyle factors, such as smoking status, alcohol intake, physical activity, annual household income, and family history of diabetes was collected using self-administered questionnaires (Supplement 1). Smoking status was categorized as: never smokers (never smoked before); former smokers (smoked ≥100 cigarettes in the past but quit before the study); and current smokers (active smokers at the time of the study). Alcohol intake was classified as follows: rare (<twice per month); sometimes (twice per month to twice per week); and often (>3 times per week). Physical activity was classified as: rare (less than once per week on average); sometimes (1 to 4 times per week); or often (>5 times per week). Annual household income was divided into three groups: low (0%–30%); middle (31%–69%); and high (70%–100%). Family history of diabetes was classified as “yes” or “no” based on the self-report questionnaire checked by the health screening examinee at baseline. The use of statins was categorized as “yes” or “no” depending on statin prescription at baseline. All covariates mentioned above were assessed on the second examination date, which served as the baseline for this study.
Statistical analysis
The statistical summary is presented as mean±standard deviation for continuous variables and percentage (%) for categorical variables. For group comparisons of baseline characteristics, analysis of variance (i.e., “ANOVA”) for continuous variables and the chi-square test for categorical variables were performed.
Kaplan-Meier estimates and log-rank tests were used to compare the cumulative incidence of diabetes with change(s) in SBP. Cox proportional hazards (Cox-PH) regression models with Pvalues and 95% confidence intervals (CIs) were used to account for confounding effects on the occurrence of diabetes. To minimize the effects of confounding variables, three Cox-PH models were applied sequentially according to the following confounding variables: Model 1, unadjusted; Model 2, adjusted for age, BMI, smoking status, alcohol intake, physical activity, and annual household income; and Model 3, fully adjusted for serum fasting glucose, total cholesterol, alanine transaminase, family history of diabetes, and statin(s) prescription, in addition to the variables in Model 2. Additionally, a fully adjusted Cox-PH model was applied to investigate the association between changes in SBP and diabetes risk after stratifying the population into normotensive (<120 mm Hg), prehypertensive (120–139 mm Hg), and hypertensive (≥140 mm Hg) according to SBP levels at baseline.
All statistical tests were two-sided, and differences with P-value <0.05 were considered to be significant. Statistical analyses were performed using SAS Enterprise Guide ver. 7.1 (SAS Inc.) and R studio ver. 3.3.3 (R Core Team, The R Foundation for Statistical Computing).
Results
Participant characteristics stratified according to sex
The median follow-up duration was 14.3 years. Characteristics of the participants according to changes in SBP stratified according to sex are summarized in Table 1. Of 152,547 participants, 87,860 (57.6%) were male and 64,687 (42.4%) were female. The three groups, categorized as decrease, no change, and increase according to changes in SBP, included 26,852, 32,960, and 28,048 males and 20,221, 24,403, and 20,063 females, respectively. For both sexes, the no-change group had the youngest participants whereas the increase group had the oldest (51.9, 51.2, and 52.2 years for males and 53.5, 52.3, and 53.6 years for females in the decrease, no change, and increase groups, respectively). At baseline, SBP and fasting blood glucose levels were highest in the decrease group and lowest in the increase group for both sexes. Among males, the proportion of current smokers and often drinkers was the highest in the increase group. Among females, smoking status was not significantly different among the three groups, and the proportion of often drinkers was slightly higher in the decrease and increase groups than that in the no-change group. The number of participants with a family history of diabetes was the highest in the no-change group in both sexes.
Based on baseline SBP in males, the hypertensive group was older than the other groups and had the highest BMI, fasting blood glucose, and total cholesterol levels among the three groups. It most likely comprised current smokers and often drinkers (Supplement 2). Similar patterns were observed among females; however, smoking status alone was higher in the normotensive group than that in the prehypertensive or hypertensive groups.
Estimated cumulative incidence of diabetes according to changes in SBP
The estimated cumulative incidence of diabetes according to change(s) in SBP during follow-up, stratified according to sex, is reported in Figure 2. Diabetes occurred in 26,532 participants (male, n=18,444; female, n=8,088) during the follow-up period. The estimated cumulative incidence of diabetes was lowest in the no-change group in both males and females. Among males, the increase group exhibited the highest cumulative incidence of diabetes, whereas in females, there was no significant difference between the decrease and increase groups.
Hazard ratios for new-onset diabetes according to changes in SBP
Hazard ratios (HRs) for new-onset diabetes according to change(s) in SBP are summarized in Table 2. In Model 1, compared with the no-change group, the HRs for the decrease and increase groups were 1.09 (95% CI, 1.06–1.13) and 1.12 (95% CI, 1.08–1.20) in males and 1.14 (95% CI, 1.08–1.20) and 1.14 (95% CI, 1.08–1.20) in females, respectively. In Model 2, the HRs for the decrease and increase groups after adjustment were 1.06 (95% CI, 1.02–1.10) and 1.10 (95% CI, 1.06–1.14) for males, and 1.06 (95% CI, 1.00–1.12) and 1.08 (95% CI, 1.02–1.13) for females, respectively. Finally, the HRs, after full adjustment for variables in Model 3, in addition to those in Model 2, were 1.06 (95% CI, 1.02–1.10) and 1.10 (95% CI, 1.06–1.14) for males, and 1.06 (95% CI, 1.00–1.12) and 1.08 (95% CI, 1.02–1.13) for females in the decrease and increase groups, respectively.
Considering the possibility that the risk for diabetes may vary depending on baseline SBP, Cox-PH models were constructed after stratifying baseline SBP into normotensive, prehypertensive, and hypertensive groups (Table 3). Among males, the HRs for the decrease and increase groups compared with the no-change group were 1.16 (95% CI, 1.08–1.24) and 1.01 (95% CI, 0.89–1.15) in the normotensive group, 1.14 (95% CI, 1.09–1.20) and 0.93 (95% CI, 0.88–0.97) in the prehypertensive group, and 1.08 (95% CI, 0.98–1.18) and 0.99 (95% CI, 0.93–1.05) in the hypertensive group, respectively. Among females, the HRs were 1.11 (95% CI, 1.02–1.21) and 0.88 (95% CI, 0.76–1.02) in the normotensive group, 1.19 (95% CI, 1.10–1.29) and 0.96 (95% CI, 0.89–1.03) in the prehypertensive group, and 1.08 (95% CI, 0.93–1.24) and 0.92 (95% CI, 0.83–1.02) in the hypertensive group, respectively.
Discussion
This study found that changes in SBP—particularly a decrease—in normotensive or prehypertensive individuals who were not taking antihypertensive drugs at baseline were associated with a high risk for diabetes.
Several studies have reported that increased SBP is a risk factor for diabetes. A meta-analysis revealed that an increase in SBP by 1 mm Hg increased the risk for type 2 diabetes by 2% [4], while another reported that a 20 mm Hg increase in SBP was associated with a 58% higher risk of diabetes [5]. These findings are consistent with those reported in other Asian countries. A large prospective cohort study involving older Chinese adults demonstrated that increased BP was associated with a higher incidence of type 2 diabetes than was normal BP [10]. A study involving 7,150 middle-age Koreans revealed a 27% and 51% higher risk for type 2 diabetes with prehypertension and hypertension, respectively [11].
Type 2 diabetes and hypertension are closely related because they have similar risk factors including endothelial dysfunction, vascular inflammation, arterial remodeling, atherosclerosis, dyslipidemia, and obesity [12,13]. However, the mechanisms underlying this association remain unclear. One potential underlying mechanism is insulin resistance, which plays a key role in the occurrence of type 2 diabetes but may also cause other cardiovascular diseases including hypertension [14]. Insulin induces vasodilation through nitric oxide production in the vascular system [15,16]. However, resistance impairs both these processes and contributes to increased BP [17]. In hypertension, the renin-angiotensin system, sympathetic nervous system, and oxidative stress are the pathophysiological factors involved in insulin resistance. The dysregulation of the renin-angiotensin system in hypertension contributes to insulin resistance via signal dysfunction [17,18]. Based on several proposed mechanisms, it can be inferred that an increased BP is a risk factor for diabetes.
As previously reported, hypertension and diabetes may be interrelated, and commonly coexist [19]. Collectively, these are substantial risk factors for coronary artery disease, cerebrovascular disease, and renal failure [19]. Therefore, strict control of BP and glucose levels is required to prevent cardiovascular complications. Several studies have shown that it is beneficial in controlling BP in patients with diabetes [20-22]. However, few studies have examined these effects in individuals without diabetes. This study was conducted in Korea and included 2,225 participants. It was reported that baseline SBP was not significantly related to diabetes; however, an increase in SBP was independently related to the incidence of diabetes [7]. In contrast, some studies have reported a decrease in SBP and diabetes; however, the results have been inconsistent. A study based on the results of the Systolic Blood Pressure Intervention Trial Randomized Trials found that intensive control of SBP to <120 mm Hg in patients with hypertension without diabetes was not associated with a risk for developing new-onset diabetes [8]. However, a meta-analysis of 22 studies reported that lowering SBP by 5 mm Hg reduced the risk for type 2 diabetes by 11% [6].
Because previous studies have shown that decreased SBP reduces or is unrelated to the risk of diabetes, the mechanism(s) that explain the correlation between increased SBP and diabetes risk alone would have been sufficient to support these results. However, they cannot sufficiently explain the findings of our study, which attempted to clarify inconsistencies in previous investigations. Unlike previous investigations, a decrease in SBP increased the risk for diabetes in this study. This result could probably be explained by activity of the autonomic nervous system. A decrease in BP causes an increase in sympathetic outflow and the suppression of vagal nerve activity as compensatory action [23]. Activation of the sympathetic nervous system increases glucose levels by inhibiting insulin secretion from pancreatic β-cells [24]. It is also involved in hepatic glucose production through direct stimulation of sympathetic nerves in the liver, epinephrine in the adrenal medulla, and pancreatic glucagon [24]. Although these responses are primarily short- to mid-term physiological reactions, several studies have suggested that chronic sympathetic over-activity may also contribute to long-term metabolic disturbances, such as insulin resistance and pancreatic β-cell dysfunction, which are critical in the development of type 2 diabetes over time [25,26]. Furthermore, long-term BP variability, which may reflect autonomic dysfunction, has been associated with an increased risk for incident diabetes, independent of mean BP levels [27]. These findings support the plausibility of autonomic dysregulation, which may be a sustained contributor to the observed association between decreased SBP and increased risk for diabetes over a long follow-up.
Strengths and limitations
Our study had several strengths. First, the effect of antihypertensive drugs on changes in SBP was mitigated by excluding participants taking these drugs. Baseline SBP was stratified into normotensive, prehypertensive, and hypertensive SBP levels, and the risk for diabetes according to changes in SBP was calculated for each level. Second, this study used data representative of the entire Korean population. Most Koreans have obligatory national health insurance run by the NHIS, which conducts a biennial national health screening program for subscribers ≥40 years of age. Thus, it was possible to include a relatively large sample size. Third, because most of the population is registered in this system, the possibility of missing or biased records was minimized. Fourth, considering that chronic diseases, such as diabetes, do not typically occur within a short period, our study had a relatively long follow-up (median, 14.3 years), enabling a more reliable assessment of diabetes occurrence. Fifth, the possibility of distortion of the results due to diagnostic errors was low because diabetes diagnosis and antidiabetic drug administration were checked using health insurance claims and laboratory data.
However, the present study also had several limitations. First, the reason for the observed decrease in SBP compared with baseline remains unclear. Therefore, it could not be determined whether the decrease was the result or the cause of diabetes. For example, weight loss before the diagnosis of diabetes due to pancreatic β-cell dysfunction may lead to reduced BP. Second, although we attempted to minimize the effects of confounding factors, we could not control them completely. Among the lifestyle-related risk factors for diabetes, detailed information, such as sedentary time, was not available in the dataset [28]. Therefore, we relied on physical activity level as a proxy. Third, in defining changes in BP, the interval between the initial and subsequent BP measurements was limited to a maximum of 3 years. Although we intended to secure a long follow-up period, considering that chronic diseases, such as diabetes, do not develop within a short time, the interval was limited to avoid affecting the results due to variations in the length of the follow-up periods. Increased BP is a common feature of aging [29]; accordingly, the prevalence of hypertension is higher in older adults. Considering that diabetes also increases in prevalence with age, some participants included in this study may have been relatively young at both the start and end of the follow-up; therefore, the possibility that this affected the analysis results cannot be ruled out. Fourth, it was difficult to evaluate and correct for general conditions, such as dehydration caused by long fasting or psychological tension, on the day of health screening, which could have affected the measurements. In addition, causes of changes in BP measured in hospitals and homes, such as “white-coat” or masked hypertension, could not be controlled and may have affected the actual risk assessment. Finally, the relatively long median follow-up (14.3 years) may have introduced residual confounding factors due to unmeasured, time-varying factors. Changes in lifestyle behaviors, such as diet, smoking, alcohol consumption, body weight, or the development of subclinical health conditions over time, may have influenced both SBP and diabetes risk, but were not fully captured by the baseline assessment or in the limited follow-up data.
Conclusion
In conclusion, changes in SBP (increase or decrease) were associated with a higher risk of diabetes. However, paradoxically, the risk for diabetes increased when the initial SBP corresponded to normal or prehypertension and, simultaneously, the next measured SBP decreased. Based on these findings, there was no consistent direction explaining the association between changes in SBP and the incidence of diabetes; as such, further research aiming to confirm these changes as risk factors for diabetes is warranted.
Notes
Conflict of interest
No potential conflict of interest relevant to this article was reported.
Funding
None.
Data availability
The data used in this study are from the Korean National Health Insurance Service (NHISS) data-sharing service. NHISS data is accessible only to authorized users. To access the data, when users complete an application for research purposes, NHISS, the data provider, reviews it and permits access to the database.
Author contribution
Conceptualization: HSY, HTK. Data curation: JSK, JK. Formal analysis: JSK, JK. Investigation: all authors. Methodology: HSY, HTK. Supervision: HTK, JK. Writing–original draft: HSY. Writing–review & editing: JK, HTK. Final approval of the manuscript: all authors.
Flow-diagram illustrating the participant selection process and exclusion criteria. ICD-10, International Classification of Diseases, 10th Revision.
Figure. 2.
Estimated cumulative incidence of diabetes mellitus and changes in systolic blood pressure. A comparison of the cumulative incidence of diabetes during follow-up according to changes in systolic blood pressure and sex using Kaplan-Meier estimates and log-rank tests.
Table 1.
Participant characteristics according to changes in systolic blood pressurea)
Characteristic
Decrease
No change
Increase
P-value
Male (n=87,860)
No. of patients
26,852
32,960
28,048
Age (y)
51.9±8.2
51.2±7.8
52.2±8.3
<0.001
Body mass index (kg/m2)
23.9±2.8
23.7±2.7
23.8±2.8
<0.001
Systolic blood pressure (mm Hg)
135.2±15.7
124.4±13.0
117.9±13.6
<0.001
Fasting blood glucose (mg/dL)
92.5±12.5
92.2±12.2
91.8±12.5
<0.001
Total cholesterol (mg/dL)
197.1±35.7
195.8±34.8
195.2±35.1
<0.001
Alanine transaminase (mg/dL)
28.3±19.8
27.7±19.7
27.7±21.5
<0.001
Smoking status
<0.001
Never smokers
12,320 (45.9)
14,888 (45.2)
12,942 (46.1)
Former smokers
4,359 (16.2)
5,610 (17.0)
4,232 (15.1)
Current smokers
10,173 (37.9)
12,462 (37.8)
10,874 (38.8)
Alcohol intake
<0.001
Rare
9,177 (34.2)
11,612 (35.2)
9,967 (35.5)
Sometimes
12,848 (47.8)
16,215 (49.2)
13,385 (47.7)
Often
4,827 (18.0)
5,133 (15.6)
4,696 (16.7)
Physical activity
<0.001
Rare
11,627 (43.6)
13,740 (41.7)
12,171 (43.4)
Sometimes
12,771 (47.6)
16,202 (49.2)
13,124 (46.8)
Often
2,382 (8.9)
3,018 (9.2)
2,752 (9.8)
Household income
<0.001
Low (0%–30%)
3,977 (14.8)
4,134 (12.5)
4,326 (15.4)
Middle (31%–69%)
5,360 (20.0)
6,349 (19.3)
5,850 (20.9)
High (70%–100%)
17,515 (65.2)
22,477 (68.2)
17,872 (63.7)
Family history of diabetes
1,344 (5.0)
1,862 (5.6)
1,374 (4.9)
<0.001
Statin use
1,052 (3.9)
1,236 (3.8)
1,328 (4.7)
<0.001
Female (n=64,687)
No. of patients
20,221
24,403
20,063
Age (y)
53.5±8.7
52.3±8.2
53.6±8.8
<0.001
Body mass index (kg/m2)
23.6±3.0
23.4±2.8
23.5±3.0
<0.001
Systolic blood pressure (mm Hg)
131.2±16.9
119.7±14.1
113.6±14.5
<0.001
Fasting blood glucose (mg/dL)
89.9±11.5
89.7±11.3
89.6±11.5
0.033
Total cholesterol (mg/dL)
201.1±36.8
198.3±36.2
198.9±36.4
<0.001
Alanine transaminase (mg/dL)
20.4±13.1
19.9±14.4
20.2±15.6
<0.001
Smoking status
0.314
Never smokers
19,753 (97.7)
23,859 (97.8)
19,617 (97.8)
Former smokers
130 (0.6)
179 (0.7)
124 (0.6)
Current smokers
338 (1.7)
365 (1.5)
322 (1.6)
Alcohol intake
0.030
Rare
16,592 (82.1)
19,814 (81.2)
16,476 (82.1)
Sometimes
3,326 (16.4)
4,242 (17.4)
3,285 (16.4)
Often
303 (1.5)
347 (1.4)
302 (1.5)
Physical activity
<0.001
Rare
12,329 (61.0)
13,834 (56.7)
11,757 (58.6)
Sometimes
6,055 (29.9)
8,207 (33.6)
6,256 (31.2)
Regular
1,837 (9.1)
2,362 (9.7)
2,050 (10.2)
Household income
<0.001
Low (0%–30%)
5,610 (27.7)
6,575 (26.9)
5,719 (28.5)
Middle (31%–69%)
5,092 (25.2)
5,705 (23.4)
5,011 (25.0)
High (70%–100%)
9,519 (47.1)
12,123 (49.7)
9,333 (46.5)
Family history of diabetes
1,245 (6.2)
1,698 (7.0)
1,248 (6.2)
<0.001
Statin use
1,295 (6.4)
1,409 (5.8)
1,417 (7.1)
0.001
Values are presented as mean±standard deviation or number (%).
a)Changes of systolic blood pressure: decrease, systolic blood pressure decrease of ≥10 mm Hg; no change, change of <10 mm Hg; increase, systolic blood pressure increase of ≥10 mm Hg.
Table 2.
Cox proportional hazards regression models for new-onset diabetes mellitus according to changes in systolic blood pressurea)
Variable
HR (95% CI)
Decrease
No change
Increase
Male
Model 1
1.09 (1.06–1.13)
1
1.12 (1.08–1.20)
Model 2
1.06 (1.02–1.10)
1
1.10 (1.06–1.14)
Model 3
1.06 (1.02–1.10)
1
1.10 (1.06–1.14)
Female
Model 1
1.14 (1.08–1.20)
1
1.14 (1.08–1.20)
Model 2
1.06 (1.00–1.12)
1
1.08 (1.02–1.13)
Model 3
1.06 (1.00–1.12)
1
1.08 (1.02–1.13)
Model 1: unadjusted; Model 2: adjusted for age, body mass index, smoking status, alcohol intake, physical activity, and household income; Model 3: adjusted for levels of serum fasting glucose, total cholesterol, alanine transaminse, family history of diabetes mellitus, and statin prescription, in addition to the variables in Model 2.
HR, hazard ratio; CI, confidence interval.
a)Changes of systolic blood pressure: decrease, systolic blood pressure decrease of ≥10 mm Hg; no change, change of <10 mm Hg; increase, systolic blood pressure increase of ≥10 mm Hg.
Table 3.
HRs for new-onset diabetes mellitus according to changes in SBP from baselinea)
a)Changes of SBP: decrease, SBP decrease of ≥10 mm Hg; no change, change of <10 mm Hg; increase, SBP increase of ≥10 mm Hg.
b)Normotensive, prehypertensive, and hypertensive group were classified according to SBP at baseline and defined as <120 mm Hg, 120–139 mm Hg, and ≥140 mm Hg, respectively.
2. Bellou V, Belbasis L, Tzoulaki I, Evangelou E. Risk factors for type 2 diabetes mellitus: an exposure-wide umbrella review of meta-analyses. PLoS One 2018;13:e0194127.
3. Nabila S, Kim JE, Choi J, Park J, Shin A, Lee SA, et al. Associations between modifiable risk factors and changes in glycemic status among individuals with prediabetes. Diabetes Care 2023;46:535-43.
4. Aikens RC, Zhao W, Saleheen D, Reilly MP, Epstein SE, Tikkanen E, et al. Systolic blood pressure and risk of type 2 diabetes: a mendelian randomization study. Diabetes 2017;66:543-50.
5. Emdin CA, Anderson SG, Woodward M, Rahimi K. Usual blood pressure and risk of new-onset diabetes: evidence from 4.1 million adults and a meta-analysis of prospective studies. J Am Coll Cardiol 2015;66:1552-62.
6. Nazarzadeh M, Bidel Z, Canoy D, Copland E, Wamil M, Majert J, et al. Blood pressure lowering and risk of new-onset type 2 diabetes: an individual participant data meta-analysis. Lancet 2021;398:1803-10.
7. Lee SW, Kim HC, Lee JM, Yun YM, Lee JY, Suh I. Association between changes in systolic blood pressure and incident diabetes in a community-based cohort study in Korea. Hypertens Res 2017;40:710-6.
8. Roumie CL, Hung AM, Russell GB, Basile J, Kreider KE, Nord J, et al. Blood pressure control and the association with diabetes mellitus incidence: results from SPRINT randomized trial. Hypertension 2020;75:331-8.
9. National Health Insurance Sharing Service. Guide for health checkup cohort database [Internet]. National Health Insurance Sharing Service; 2025 [cited 2025 May 9]. Available from: https://nhiss.nhis.or.kr/bd/ab/bdaba006cv.do#
10. Yang X, Chen J, Pan A, Wu JH, Zhao F, Xie Y, et al. Association between higher blood pressure and risk of diabetes mellitus in middle-aged and elderly Chinese adults. Diabetes Metab J 2020;44:436-45.
11. Kim MJ, Lim NK, Choi SJ, Park HY. Hypertension is an independent risk factor for type 2 diabetes: the Korean genome and epidemiology study. Hypertens Res 2015;38:783-9.
14. Ormazabal V, Nair S, Elfeky O, Aguayo C, Salomon C, Zuniga FA. Association between insulin resistance and the development of cardiovascular disease. Cardiovasc Diabetol 2018;17:122.
17. Mancusi C, Izzo R, di Gioia G, Losi MA, Barbato E, Morisco C. Insulin resistance the hinge between hypertension and type 2 diabetes. High Blood Press Cardiovasc Prev 2020;27:515-26.
21. Knowler WC, Bennett PH, Ballintine EJ. Increased incidence of retinopathy in diabetics with elevated blood pressure: a six-year follow-up study in Pima Indians. N Engl J Med 1980;302:645-50.
22. UK Prospective Diabetes Study Group. Tight blood pressure control and risk of macrovascular and microvascular complications in type 2 diabetes: UKPDS 38. UK Prospective Diabetes Study Group. BMJ 1998;317:703-13.
27. Zhou R, Li FR, Liu K, Huang RD, Liu HM, Yuan ZL, et al. Long-term visit-to-visit blood pressure variability and risk of diabetes mellitus in Chinese population: a retrospective population-based study. Int J Public Health 2023;68:1605445.
28. Patterson R, McNamara E, Tainio M, de Sa TH, Smith AD, Sharp SJ, et al. Sedentary behaviour and risk of all-cause, cardiovascular and cancer mortality, and incident type 2 diabetes: a systematic review and dose response meta-analysis. Eur J Epidemiol 2018;33:811-29.
29. Gurven M, Blackwell AD, Rodríguez DE, Stieglitz J, Kaplan H. Does blood pressure inevitably rise with age?: longitudinal evidence among forager-horticulturalists. Hypertension 2012;60:25-33.
Association between changes in systolic blood pressure and the incidence of diabetes mellitus: a retrospective study based on the Korean National Health Screening Cohort
Figure. 1. Flow-diagram illustrating the participant selection process and exclusion criteria. ICD-10, International Classification of Diseases, 10th Revision.
Figure. 2. Estimated cumulative incidence of diabetes mellitus and changes in systolic blood pressure. A comparison of the cumulative incidence of diabetes during follow-up according to changes in systolic blood pressure and sex using Kaplan-Meier estimates and log-rank tests.
Graphical abstract
Figure. 1.
Figure. 2.
Graphical abstract
Association between changes in systolic blood pressure and the incidence of diabetes mellitus: a retrospective study based on the Korean National Health Screening Cohort
Characteristic
Decrease
No change
Increase
P-value
Male (n=87,860)
No. of patients
26,852
32,960
28,048
Age (y)
51.9±8.2
51.2±7.8
52.2±8.3
<0.001
Body mass index (kg/m2)
23.9±2.8
23.7±2.7
23.8±2.8
<0.001
Systolic blood pressure (mm Hg)
135.2±15.7
124.4±13.0
117.9±13.6
<0.001
Fasting blood glucose (mg/dL)
92.5±12.5
92.2±12.2
91.8±12.5
<0.001
Total cholesterol (mg/dL)
197.1±35.7
195.8±34.8
195.2±35.1
<0.001
Alanine transaminase (mg/dL)
28.3±19.8
27.7±19.7
27.7±21.5
<0.001
Smoking status
<0.001
Never smokers
12,320 (45.9)
14,888 (45.2)
12,942 (46.1)
Former smokers
4,359 (16.2)
5,610 (17.0)
4,232 (15.1)
Current smokers
10,173 (37.9)
12,462 (37.8)
10,874 (38.8)
Alcohol intake
<0.001
Rare
9,177 (34.2)
11,612 (35.2)
9,967 (35.5)
Sometimes
12,848 (47.8)
16,215 (49.2)
13,385 (47.7)
Often
4,827 (18.0)
5,133 (15.6)
4,696 (16.7)
Physical activity
<0.001
Rare
11,627 (43.6)
13,740 (41.7)
12,171 (43.4)
Sometimes
12,771 (47.6)
16,202 (49.2)
13,124 (46.8)
Often
2,382 (8.9)
3,018 (9.2)
2,752 (9.8)
Household income
<0.001
Low (0%–30%)
3,977 (14.8)
4,134 (12.5)
4,326 (15.4)
Middle (31%–69%)
5,360 (20.0)
6,349 (19.3)
5,850 (20.9)
High (70%–100%)
17,515 (65.2)
22,477 (68.2)
17,872 (63.7)
Family history of diabetes
1,344 (5.0)
1,862 (5.6)
1,374 (4.9)
<0.001
Statin use
1,052 (3.9)
1,236 (3.8)
1,328 (4.7)
<0.001
Female (n=64,687)
No. of patients
20,221
24,403
20,063
Age (y)
53.5±8.7
52.3±8.2
53.6±8.8
<0.001
Body mass index (kg/m2)
23.6±3.0
23.4±2.8
23.5±3.0
<0.001
Systolic blood pressure (mm Hg)
131.2±16.9
119.7±14.1
113.6±14.5
<0.001
Fasting blood glucose (mg/dL)
89.9±11.5
89.7±11.3
89.6±11.5
0.033
Total cholesterol (mg/dL)
201.1±36.8
198.3±36.2
198.9±36.4
<0.001
Alanine transaminase (mg/dL)
20.4±13.1
19.9±14.4
20.2±15.6
<0.001
Smoking status
0.314
Never smokers
19,753 (97.7)
23,859 (97.8)
19,617 (97.8)
Former smokers
130 (0.6)
179 (0.7)
124 (0.6)
Current smokers
338 (1.7)
365 (1.5)
322 (1.6)
Alcohol intake
0.030
Rare
16,592 (82.1)
19,814 (81.2)
16,476 (82.1)
Sometimes
3,326 (16.4)
4,242 (17.4)
3,285 (16.4)
Often
303 (1.5)
347 (1.4)
302 (1.5)
Physical activity
<0.001
Rare
12,329 (61.0)
13,834 (56.7)
11,757 (58.6)
Sometimes
6,055 (29.9)
8,207 (33.6)
6,256 (31.2)
Regular
1,837 (9.1)
2,362 (9.7)
2,050 (10.2)
Household income
<0.001
Low (0%–30%)
5,610 (27.7)
6,575 (26.9)
5,719 (28.5)
Middle (31%–69%)
5,092 (25.2)
5,705 (23.4)
5,011 (25.0)
High (70%–100%)
9,519 (47.1)
12,123 (49.7)
9,333 (46.5)
Family history of diabetes
1,245 (6.2)
1,698 (7.0)
1,248 (6.2)
<0.001
Statin use
1,295 (6.4)
1,409 (5.8)
1,417 (7.1)
0.001
Variable
HR (95% CI)
Decrease
No change
Increase
Male
Model 1
1.09 (1.06–1.13)
1
1.12 (1.08–1.20)
Model 2
1.06 (1.02–1.10)
1
1.10 (1.06–1.14)
Model 3
1.06 (1.02–1.10)
1
1.10 (1.06–1.14)
Female
Model 1
1.14 (1.08–1.20)
1
1.14 (1.08–1.20)
Model 2
1.06 (1.00–1.12)
1
1.08 (1.02–1.13)
Model 3
1.06 (1.00–1.12)
1
1.08 (1.02–1.13)
SBP at baselineb)
HR (95% CI)
Decrease
No change
Increase
Male
Normotensive
1.16 (1.08–1.24)
1
1.01 (0.89–1.15)
Prehypertensive
1.14 (1.09–1.20)
1
0.93 (0.88–0.97)
Hypertensive
1.08 (0.98–1.18)
1
0.99 (0.93–1.05)
Female
Normotensive
1.11 (1.02–1.21)
1
0.88 (0.76–1.02)
Prehypertensive
1.19 (1.10–1.29)
1
0.96 (0.89–1.03)
Hypertensive
1.08 (0.93–1.24)
1
0.92 (0.83–1.02)
Table 1. Participant characteristics according to changes in systolic blood pressurea)
Values are presented as mean±standard deviation or number (%).
Changes of systolic blood pressure: decrease, systolic blood pressure decrease of ≥10 mm Hg; no change, change of <10 mm Hg; increase, systolic blood pressure increase of ≥10 mm Hg.
Table 2. Cox proportional hazards regression models for new-onset diabetes mellitus according to changes in systolic blood pressurea)
Model 1: unadjusted; Model 2: adjusted for age, body mass index, smoking status, alcohol intake, physical activity, and household income; Model 3: adjusted for levels of serum fasting glucose, total cholesterol, alanine transaminse, family history of diabetes mellitus, and statin prescription, in addition to the variables in Model 2.
HR, hazard ratio; CI, confidence interval.
Changes of systolic blood pressure: decrease, systolic blood pressure decrease of ≥10 mm Hg; no change, change of <10 mm Hg; increase, systolic blood pressure increase of ≥10 mm Hg.
Table 3. HRs for new-onset diabetes mellitus according to changes in SBP from baselinea)
Changes of SBP: decrease, SBP decrease of ≥10 mm Hg; no change, change of <10 mm Hg; increase, SBP increase of ≥10 mm Hg.
Normotensive, prehypertensive, and hypertensive group were classified according to SBP at baseline and defined as <120 mm Hg, 120–139 mm Hg, and ≥140 mm Hg, respectively.
