https://orcid.org/0000-0003-2553-6170
Abstract
Isolated diastolic hypertension (IDH) is a largely underrated risk factor for cardiovascular disease (CVD). It is currently unclear whether a reduction in blood pressure (BP) is associated with CVD events among adults with IDH. We aimed to elucidate the relationship between BP reduction and incident CVD in individuals with IDH.
We retrospectively analysed the data of 71 297 individuals with IDH. Isolated diastolic hypertension was defined as systolic BP of < 140 mmHg and diastolic BP (DBP) of ≥90 mmHg (median age, 48 years; 83.1% men; median DBP, 92 mmHg). None of the participants took BP-lowering medications or had a history of CVD at baseline. Blood pressure was measured at baseline and 1-year follow-up, and participants were categorized into two groups based on DBP at 1 year (≥90 or < 90 mmHg). The primary outcome was a composite endpoint that included myocardial infarction, stroke, and all-cause death. Over a mean follow-up period of 1100 ± 859 days, 1317 composite CVD endpoints were recorded. Participants with DBP of < 90 mmHg at 1 year were at a lower risk of composite CVD events [hazard ratio (HR): 0.75, 95% confidence interval (CI): 0.67–0.83] than those with DBP of ≥90 mmHg at 1 year. A reduction in DBP per 5 mmHg during the 1-year follow-up was associated with a lower composite CVD event risk (HR: 0.92, 95% CI: 0.89–0.95). The results remained consistent across a multitude of sensitivity analyses.
Our analysis of a large-scale epidemiological dataset demonstrated a relationship of reduction in DBP with a reduced risk for CVD events in individuals with IDH.
Lay Summary
In people with isolated diastolic hypertension (IDH), a reduction of diastolic blood pressure (BP) was associated with a reduced risk for developing cardiovascular disease. Our results suggest a potential benefit of BP reduction in IDH and provide novel knowledge in preventive cardiology.
Lay Summary
In people with isolated diastolic hypertension (IDH), a reduction of diastolic blood pressure (BP) was associated with a reduced risk for developing cardiovascular disease.
Our results suggest a potential benefit of BP reduction in IDH and provide novel knowledge in preventive cardiology.
Introduction
Isolated diastolic hypertension (IDH), a subtype of hypertension, is defined as diastolic blood pressure (DBP) in the hypertensive range with systolic blood pressure (SBP) outside the hypertensive range. It is a relatively common condition, particularly in younger people. Previous studies on cardiovascular outcomes associated with IDH have shown varying results,1–6 and the clinical significance of IDH remains debatable. To address this long-standing debate, examining whether a reduction in DBP is associated with improved cardiovascular outcomes in patients with IDH would be helpful. However, available literature contains limited data on the relationship between changes in blood pressure (BP) and cardiovascular outcomes in people with IDH. This study aimed to clarify whether a reduction in DBP could decrease the risk of cardiovascular disease (CVD) development in people with IDH. We analysed a nationwide health check-up and administrative claims dataset.
Methods
Study population
The JMDC Claims Database (JMDC Inc., Tokyo, Japan), a health check-up and administrative claims database, was used in the present study, which is a retrospective observational cohort study.6,7 Most of those registered in the database were employees of a relatively sizeable Japanese company covered by ‘kempo’, a health insurance programme for employees. This database contains data on registered individuals’ health, including BP, body mass index (BMI), laboratory data, and responses to a self-reported questionnaire on lifestyle (e.g. cigarette smoking). Additionally, this database contains administrative claims records, and the diagnosis of CVD events is recorded using the International Classification of Diseases, 10th Revision (ICD-10) coding. We extracted from the JMDC Claims Database data on adults who underwent a health check-up that included a BP examination more than a year after insurance enrolment (1-year look-back period), with available BP data at 1 year after (see Supplementary material online, Figure S1). This cohort study included individuals with IDH, which was defined as SBP < 140 mmHg and DBP ≥90 mmHg according to the 2018 European Society of Cardiology and the European Society of Hypertension guidelines at the initial health check-up.8 We excluded individuals with a hitory of CVD, including myocardial infarction (MI), angina pectoris, stroke, and heart failure (HF); those with a history of renal replacement therapy, including haemodialysis, peritoneal dialysis, and kidney transplantation; those taking any BP-lowering medications at the initial health check-up; those who developed CVD within 1 year after the initial health check-up; those with missing data on covariates such as cigarette smoking; and those taking any BP-lowering medications at 1 year after the initial health check-up (see Supplementary material online, Figure S2).
Ethics
The Clinical Research Review Board of the University of Tokyo approved this study (number: 2018-10862), and it was conducted in compliance with the Declaration of Helsinki. The requirement for obtaining informed consent from each study participant was waived because all the data recorded in the JMDC Claims Database were anonymized and deidentified. The JMDC Claims Database is available for anyone who purchases this dataset from JMDC Inc. (https://www.eng.phm-jmdc.com/), a medical venture company in Japan.
Measurements and definitions
Health check-up data were collected using standardized health check-up protocols, and the following information was obtained: SBP, DBP, BMI, fasting laboratory data, and responses to a self-reported questionnaire on lifestyle. Trained healthcare professionals (e.g. nurses) measured the individuals’ BP at least twice after rest, and the average BP values were recorded.8 Detailed methods of BP measurement are presented in the Supplementary material online. Isolated diastolic hypertension was defined as SBP < 140 mmHg and DBP ≥90 mmHg.8 We defined an overweight status or obesity as a BMI of ≥25 kg/m2. Individuals were defined as having diabetes mellitus if their fasting glucose levels were ≥126 mg/dL or if they were using glucose-lowering medications. Dyslipidaemia was defined as a low-density lipoprotein cholesterol level of ≥140 mg/dL, a high-density lipoprotein cholesterol level of < 40 mg/dL, a triglyceride level of ≥150 mg/dL or use of lipid-lowering medications.9 Self-reported cigarette smoking status (current or non-current) was obtained.
Outcomes
We collected outcome data that occurred between January 2005 and April 2021. We defined a composite endpoint that included MI, stroke, and all-cause death, as the primary outcome. We also analysed MI, stroke, all-cause death, and HF, separately, as secondary outcomes. The Supplementary material online provides a summary of the ICD-10 codes used in the present study.
Statistical analysis
Continuous variables are shown as medians (interquartile range), and categorical variables are shown as numbers (percentages). We categorized study participants based on their DBP 1 year after their initial health check-ups (DBP < 90 mmHg vs. DBP ≥90 mmHg at 1-year follow-up). Data between the two groups were compared using the Wilcoxon rank-sum and χ2 tests for continuous and categorical variables, respectively. We performed the Cox proportional hazard regression to examine the association between the DBP category at 1 year after the initial health check-up and the subsequent CVD events. We set DBP ≥90 mmHg at 1 year after the initial health check-up as a reference. Model 1 only included the DBP category at 1 year (unadjusted model). Model 2 included the DBP category at 1 year; age; sex; BMI, SBP, diabetes mellitus, dyslipidaemia; and cigarette smoking at baseline.
To confirm the robustness of the results, we conducted several sensitivity analyses. First, we examined the association between the change in DBP (as a continuous value) and subsequent CVD events. Second, we analysed the association between a reduction of ≥5 mmHg in DBP and incident CVD. Third, we assessed the association between DBP at 1 year and incident composite CVD using a restricted cubic spline model. We used three cut-off points for DBP, with the reference point set at a DBP of 90 mmHg. We fitted three cubic spline models using three, four, and five knots, respectively. We selected the model with three knots as it had the lowest Akaike’s information criterion. Fourth, we used a restricted cubic spline model to assess the change in DBP from the initial health check-up to 1 year after the initial health check-up. We used three cut-off points to evaluate changes in BP, with the reference point set at ±0 mmHg (no change in BP). Fifth, we assessed the relationship between DBP categories and incident CVD stratified by age, sex, and overweight status or obesity. Sixth, we redefined IDH as SBP < 130 mmHg and DBP ≥80 mmHg based on the 2017 American College of Cardiology (ACC)/American Heart Association (AHA) criteria.10 In this sensitivity analysis, we extracted individuals with IDH based on the ACC/AHA criteria. Individuals were categorized accordingly into two groups based on DBP at 1 year after the initial heath check-up (≥80 or < 80 mmHg). Seventh, dyslipidaemia was redefined as having a low-density lipoprotein cholesterol level of ≥140 mg/dL, a high-density lipoprotein cholesterol level of < 40 mg/dL if men or < 50 mg/dL if women, a triglyceride level of ≥150 mg/dL, or use of lipid-lowering medications. Lastly, to address the possibility of misclassifying IDH, we calculated the average BP values using BP data at the initial health check-up and 1 year after, and defined IDH as an average SBP and DBP of < 140 and ≥90 mmHg, respectively. We analysed the association between DBP < 90 mmHg at 2 years after the initial health check-up and a subsequent incidence of CVD. The calculated P-value was used to assess whether the null hypothesis was rejected for (two-tailed) values of P < 0.05. All statistical analyses were conducted using STATA software v17 (StataCorp LLC, College Station, TX, USA).
Results
Clinical characteristics
Initially, we extracted data on 105 111 individuals with IDH from the JMDC Claims Database. Among them, we excluded those with CVD (n = 6377), a history of renal replacement therapy (n = 28), those taking any BP-lowering medications at the initial health check-up (n = 19 085), those who developed CVD within 1 year after the initial health check-up (n = 1250), those with missing data on cigarette smoking (n = 3230), and those taking any BP-lowering medications at 1 year after the initial health check-up (n = 3844). Finally, 71 297 participants were analysed in this study (see Supplementary material online, Figure S2). Table 1 displays the clinical characteristics of the study participant. Overall, the median age was 48 (range, 42–54) years, and 59 215 (83.1%) patients were men.
Clinical characteristics
| Diastolic blood pressure at 1-year follow-up among participants with isolated diastolic hypertension at the initial health check-up | |||
|---|---|---|---|
| Diastolic blood pressure ≥ 90 mmHg at 1-year follow-up (n = 34 997) | Diastolic blood pressure < 90 mmHg at 1-year follow-up (n = 36 300) | P-value | |
| Blood pressure at the initial health check-up | |||
| Diastolic blood pressure, mmHg | 93 (91–96) | 92 (90–94) | <0.001 |
| Systolic blood pressure, mmHg | 134 (130–137) | 132 (128–136) | <0.001 |
| Blood pressure at 1 year after | |||
| Diastolic blood pressure, mmHg | 95 (92–99) | 84 (80–87) | <0.001 |
| Systolic blood pressure, mmHg | 139 (133–147) | 128 (122–134) | <0.001 |
| Age, years | 48 (43–53) | 48 (42–54) | <0.001 |
| Men, n (%) | 29 702 (84.9) | 29 513 (81.3) | <0.001 |
| Body mass index, kg/m2 | 24.5 (22.3–27) | 24.2 (22–26.7) | <0.001 |
| Diabetes mellitus, n (%) | 1488 (4.3) | 1691 (4.7) | 0.009 |
| Dyslipidaemia, n (%) | 19 831 (56.7) | 20 152 (55.5) | 0.002 |
| Cigarette smoking, n (%) | 10 917 (31.2) | 10 657 (29.4) | <0.001 |
| Laboratory data | |||
| Glucose, mg/dL | 96 (89–103) | 95 (89–103) | <0.001 |
| Low-density lipoprotein cholesterol, mg/dL | 129 (108–150) | 128 (107–150) | 0.002 |
| High-density lipoprotein cholesterol, mg/dL | 57 (48–68) | 58 (49–70) | <0.001 |
| Triglycerides, mg/dL | 112 (77–167) | 109 (75–162) | <0.001 |
| Diastolic blood pressure at 1-year follow-up among participants with isolated diastolic hypertension at the initial health check-up | |||
|---|---|---|---|
| Diastolic blood pressure ≥ 90 mmHg at 1-year follow-up (n = 34 997) | Diastolic blood pressure < 90 mmHg at 1-year follow-up (n = 36 300) | P-value | |
| Blood pressure at the initial health check-up | |||
| Diastolic blood pressure, mmHg | 93 (91–96) | 92 (90–94) | <0.001 |
| Systolic blood pressure, mmHg | 134 (130–137) | 132 (128–136) | <0.001 |
| Blood pressure at 1 year after | |||
| Diastolic blood pressure, mmHg | 95 (92–99) | 84 (80–87) | <0.001 |
| Systolic blood pressure, mmHg | 139 (133–147) | 128 (122–134) | <0.001 |
| Age, years | 48 (43–53) | 48 (42–54) | <0.001 |
| Men, n (%) | 29 702 (84.9) | 29 513 (81.3) | <0.001 |
| Body mass index, kg/m2 | 24.5 (22.3–27) | 24.2 (22–26.7) | <0.001 |
| Diabetes mellitus, n (%) | 1488 (4.3) | 1691 (4.7) | 0.009 |
| Dyslipidaemia, n (%) | 19 831 (56.7) | 20 152 (55.5) | 0.002 |
| Cigarette smoking, n (%) | 10 917 (31.2) | 10 657 (29.4) | <0.001 |
| Laboratory data | |||
| Glucose, mg/dL | 96 (89–103) | 95 (89–103) | <0.001 |
| Low-density lipoprotein cholesterol, mg/dL | 129 (108–150) | 128 (107–150) | 0.002 |
| High-density lipoprotein cholesterol, mg/dL | 57 (48–68) | 58 (49–70) | <0.001 |
| Triglycerides, mg/dL | 112 (77–167) | 109 (75–162) | <0.001 |
Data are reported as medians (interquartile range) or numbers (percentage), where appropriate.
Clinical characteristics
| Diastolic blood pressure at 1-year follow-up among participants with isolated diastolic hypertension at the initial health check-up | |||
|---|---|---|---|
| Diastolic blood pressure ≥ 90 mmHg at 1-year follow-up (n = 34 997) | Diastolic blood pressure < 90 mmHg at 1-year follow-up (n = 36 300) | P-value | |
| Blood pressure at the initial health check-up | |||
| Diastolic blood pressure, mmHg | 93 (91–96) | 92 (90–94) | <0.001 |
| Systolic blood pressure, mmHg | 134 (130–137) | 132 (128–136) | <0.001 |
| Blood pressure at 1 year after | |||
| Diastolic blood pressure, mmHg | 95 (92–99) | 84 (80–87) | <0.001 |
| Systolic blood pressure, mmHg | 139 (133–147) | 128 (122–134) | <0.001 |
| Age, years | 48 (43–53) | 48 (42–54) | <0.001 |
| Men, n (%) | 29 702 (84.9) | 29 513 (81.3) | <0.001 |
| Body mass index, kg/m2 | 24.5 (22.3–27) | 24.2 (22–26.7) | <0.001 |
| Diabetes mellitus, n (%) | 1488 (4.3) | 1691 (4.7) | 0.009 |
| Dyslipidaemia, n (%) | 19 831 (56.7) | 20 152 (55.5) | 0.002 |
| Cigarette smoking, n (%) | 10 917 (31.2) | 10 657 (29.4) | <0.001 |
| Laboratory data | |||
| Glucose, mg/dL | 96 (89–103) | 95 (89–103) | <0.001 |
| Low-density lipoprotein cholesterol, mg/dL | 129 (108–150) | 128 (107–150) | 0.002 |
| High-density lipoprotein cholesterol, mg/dL | 57 (48–68) | 58 (49–70) | <0.001 |
| Triglycerides, mg/dL | 112 (77–167) | 109 (75–162) | <0.001 |
| Diastolic blood pressure at 1-year follow-up among participants with isolated diastolic hypertension at the initial health check-up | |||
|---|---|---|---|
| Diastolic blood pressure ≥ 90 mmHg at 1-year follow-up (n = 34 997) | Diastolic blood pressure < 90 mmHg at 1-year follow-up (n = 36 300) | P-value | |
| Blood pressure at the initial health check-up | |||
| Diastolic blood pressure, mmHg | 93 (91–96) | 92 (90–94) | <0.001 |
| Systolic blood pressure, mmHg | 134 (130–137) | 132 (128–136) | <0.001 |
| Blood pressure at 1 year after | |||
| Diastolic blood pressure, mmHg | 95 (92–99) | 84 (80–87) | <0.001 |
| Systolic blood pressure, mmHg | 139 (133–147) | 128 (122–134) | <0.001 |
| Age, years | 48 (43–53) | 48 (42–54) | <0.001 |
| Men, n (%) | 29 702 (84.9) | 29 513 (81.3) | <0.001 |
| Body mass index, kg/m2 | 24.5 (22.3–27) | 24.2 (22–26.7) | <0.001 |
| Diabetes mellitus, n (%) | 1488 (4.3) | 1691 (4.7) | 0.009 |
| Dyslipidaemia, n (%) | 19 831 (56.7) | 20 152 (55.5) | 0.002 |
| Cigarette smoking, n (%) | 10 917 (31.2) | 10 657 (29.4) | <0.001 |
| Laboratory data | |||
| Glucose, mg/dL | 96 (89–103) | 95 (89–103) | <0.001 |
| Low-density lipoprotein cholesterol, mg/dL | 129 (108–150) | 128 (107–150) | 0.002 |
| High-density lipoprotein cholesterol, mg/dL | 57 (48–68) | 58 (49–70) | <0.001 |
| Triglycerides, mg/dL | 112 (77–167) | 109 (75–162) | <0.001 |
Data are reported as medians (interquartile range) or numbers (percentage), where appropriate.
Diastolic blood pressure category at 1 year and risk of cardiovascular disease
During a mean follow-up of 1100 ± 859 days, 1317 composite CVD events were recorded. The cumulative incidence of composite CVD events was lower in participants with DBP < 90 mmHg at 1 year after the initial health check-up [53.4 (49.2–57.8) per 10 000 person-years] than in those with DBP ≥90 mmHg at 1 year after the initial health check-up [71.7 (66.7–77.1) per 10 000 person-years]. In event-free survival curves derived from the multivariable Cox regression analysis, the group with DBP < 90 mmHg at 1 year after the initial health check-up had a lower cumulative incidence of composite CVD events than the group with DBP ≥90 mmHg at 1 year after the initial health check-up (Figure 1). In an unadjusted model (Model 1), DBP < 90 mmHg at 1 year after the initial health check-up was similarly associated with a lower risk of developing composite CVD events compared with DBP ≥90 mmHg at 1 year after the initial health check-up [hazard ratio (HR): 0.74, 95% confidence interval (CI): 0.66–0.83]. In the multivariable Cox regression analysis (Model 2), DBP < 90 mmHg at 1 year after the initial health check-up was associated with a lower risk for composite CVD events than DBP ≥90 mmHg at 1 year after the initial health check-up (HR: 0.75, 95% CI: 0.67–0.83) (Table 2).
Event-free survival curves for the composite endpoint. Models included adjustment for age, sex, body mass index, systolic blood pressure, diabetes mellitus, dyslipidaemia, and cigarette smoking at baseline.
Association between diastolic blood pressure at 1-year after the initial health check-up and the risk for a composite endpoint
| DBP at 1 year | |||
|---|---|---|---|
| DBP ≥ 90 mmHg at 1 year (n = 34 997) | DBP < 90 mmHg at 1 year (n = 36 300) | P-value | |
| Event | 727 | 590 | |
| Incident rate | 71.7 (66.7–77.1) | 53.4 (49.2–57.8) | |
| Model 1 | 1 (Reference) | 0.74 (0.66–0.83) | <0.001a |
| Model 2 | 1 (Reference) | 0.75 (0.67–0.83) | <0.001a |
| DBP at 1 year | |||
|---|---|---|---|
| DBP ≥ 90 mmHg at 1 year (n = 34 997) | DBP < 90 mmHg at 1 year (n = 36 300) | P-value | |
| Event | 727 | 590 | |
| Incident rate | 71.7 (66.7–77.1) | 53.4 (49.2–57.8) | |
| Model 1 | 1 (Reference) | 0.74 (0.66–0.83) | <0.001a |
| Model 2 | 1 (Reference) | 0.75 (0.67–0.83) | <0.001a |
DBP, diastolic blood pressure. The incidence rate was per 10 000 person-years. Model 1 is unadjusted. Model 2 includes adjustment for age, sex, body mass index, systolic blood pressure, diabetes mellitus, dyslipidaemia, and cigarette smoking at baseline. Statistical significance was set at Bonferroni-corrected P-value <0.025 (0.05/2) based on the number of performed analyses.
Statistically significant.
Association between diastolic blood pressure at 1-year after the initial health check-up and the risk for a composite endpoint
| DBP at 1 year | |||
|---|---|---|---|
| DBP ≥ 90 mmHg at 1 year (n = 34 997) | DBP < 90 mmHg at 1 year (n = 36 300) | P-value | |
| Event | 727 | 590 | |
| Incident rate | 71.7 (66.7–77.1) | 53.4 (49.2–57.8) | |
| Model 1 | 1 (Reference) | 0.74 (0.66–0.83) | <0.001a |
| Model 2 | 1 (Reference) | 0.75 (0.67–0.83) | <0.001a |
| DBP at 1 year | |||
|---|---|---|---|
| DBP ≥ 90 mmHg at 1 year (n = 34 997) | DBP < 90 mmHg at 1 year (n = 36 300) | P-value | |
| Event | 727 | 590 | |
| Incident rate | 71.7 (66.7–77.1) | 53.4 (49.2–57.8) | |
| Model 1 | 1 (Reference) | 0.74 (0.66–0.83) | <0.001a |
| Model 2 | 1 (Reference) | 0.75 (0.67–0.83) | <0.001a |
DBP, diastolic blood pressure. The incidence rate was per 10 000 person-years. Model 1 is unadjusted. Model 2 includes adjustment for age, sex, body mass index, systolic blood pressure, diabetes mellitus, dyslipidaemia, and cigarette smoking at baseline. Statistical significance was set at Bonferroni-corrected P-value <0.025 (0.05/2) based on the number of performed analyses.
Statistically significant.
Diastolic blood pressure category at 1 year and risk of myocardial infarction, stroke, and heart failure
During the follow-up, 258 MI, 981 stroke, 130 all-cause death, and 1963 HF events were recorded. In survival curves derived from the multivariable Cox regression analysis, the cumulative incidence of MI, stroke, all-cause death, and HF was lower in those with DBP < 90 mmHg at 1 year after the initial health check-up than in those with DBP ≥90 mmHg at 1 year after the initial health check-up (see Supplementary material online, Figure S3). After multivariable adjustment, the HRs (95% CI) of DBP < 90 mmHg at 1 year after the initial health check-up for MI, stroke, all-cause death, and HF were 0.95 (0.75–1.22), 0.66 (0.58–0.75), 1.15 (0.81–1.64), and 0.72 (0.65–0.78), respectively, compared with DBP ≥90 mmHg at 1 year after the initial health check-up (see Supplementary material online, Table S1).
Sensitivity analyses
The decrease in DBP (per 5 mmHg) as a continuous value was significantly associated with a lower risk of experiencing composite CVD events, stroke, and HF (see Supplementary material online, Figure S4). Individuals with ≥5 mmHg reduction in DBP were at a lower risk of developing composite CVD events, stroke, and HF than those whose DBP did not decrease by ≥5 mmHg (see Supplementary material online, Figure S5). The cubic splines demonstrated that the risk of composite CVD events increased linearly with DBP after DBP exceeded 80–90 mmHg (Figure 2). Furthermore, we examined the relationship between the 1-year change in DBP and composite CVD events. The restricted cubic splines of the change in DBP demonstrated that the risk of developing CVD was linearly associated with the 1-year change in DBP (Figure 3). We performed subgroup analyses. In all subgroups, the DBP category at 1 year was associated with the risk of experiencing composite CVD events. The P-value for interaction between all subgroups was not significant (see Supplementary material online, Figure S6). We extracted 293 793 individuals with IDH based on the ACC/AHA criteria. In this population, DBP < 80 mmHg at 1 year was associated with a lower risk of developing CVD events compared with DBP ≥80 mmHg at 1 year (see Supplementary material online, Figure S7). Our primary results were consistent even if we changed the definition of dyslipidaemia (see Supplementary material online, Figure S8). Lastly, we analysed 40 193 individuals with an average SBP of < 140 mmHg and DBP ≥90 mmHg at the initial and 1-year health check-ups to address the possibility of misclassification of IDH and found that DBP < 90 mmHg at 2 years after the initial health check-up was associated with a lower subsequent composite CVD events risk than those with DBP ≥90 mmHg at 2 years after the initial health check-up (see Supplementary material online, Figure S9).
Risk of major cardiovascular events across diastolic blood pressure levels at 1 year; a restricted cubic spline analysis. Among participants with isolated diastolic hypertension at baseline, we assessed the association between diastolic blood pressure at 1-year follow-up and incident composite cardiovascular disease using a restricted cubic spline regression model. We used three cut-off points for diastolic blood pressure (10, 50, and 90 percentiles), with the reference point set at diastolic blood pressure of 90 mmHg. We fitted three cubic spline models using three, four, and five knots. We selected the model with three knots since it had the lowest Akaike’s information criterion. Restricted cubic spline shows the relationship of diastolic blood pressure at 1 year after with composite endpoint. Hazard ratios were adjusted for age, sex, body mass index, systolic blood pressure, diabetes mellitus, dyslipidaemia, and cigarette smoking at baseline. P for non-linearity was 0.0056. Median diastolic blood pressure at 1 year was 89 (interquartile range, 84–95) mmHg.
Risk of major cardiovascular events across diastolic blood pressure changes at 1 year; a restricted cubic spline analysis. Among participants with isolated diastolic hypertension at baseline, we assessed the association between the change in diastolic blood pressure from baseline to 1-year follow-up and incident composite cardiovascular disease using a restricted cubic spline regression model. We used three cut-off points for change in blood pressure (10, 50, and 90 percentiles), with the reference point set at ±0 mmHg (no change in diastolic blood pressure). Restricted cubic spline shows the relationship of the changes in diastolic blood pressure over 1 year with composite endpoint. Hazard ratios were adjusted for age, sex, body mass index, systolic blood pressure, diabetes mellitus, dyslipidaemia, and cigarette smoking at baseline. P for non-linearity was 0.0060. Median change in diastolic blood pressure from the initial health check-up to 1 year after the initial health check-up was −4 (interquartile range, −9 to 2) mmHg.
Discussion
This study included ∼70 000 adults with IDH, defined on the basis of having an SBP < 140 mmHg and a DBP ≥90 mmHg, who were enroled in a large-scale database in Japan. We found an association between a reduction in DBP after 1 year and a reduced risk of developing CVD. Compared with DBP ≥90 mmHg at 1 year, DBP < 90 mmHg at 1 year was associated with a decreased incidence of CVD events, including MI, stroke, and all-cause death. This association was driven mainly by a reduction in incident stroke. We validated the robustness of the results through several sensitivity analyses. To the best of our knowledge, this is the first study evaluating the association between reduced DBP and incident CVD among individuals with IDH using a large-scale real-world dataset.
Although recent trials (e.g. the SPRINT trial11 and STEP study12) have demonstrated the benefits of strict BP control in hypertension from the perspective of CVD prevention, these trials included hypertensive individuals with high SBP levels. Additionally, to our knowledge, no data showing the clinical benefit of BP-lowering treatment for people with IDH have been presented thus far. Based on this background, we believe that our study, which shows the potential benefit of BP reduction for IDH, is valuable. The large sample size of our dataset enabled a variety of sensitivity analyses, which strengthened the robustness of our results. Our primary findings were derived mainly from the risk reduction in stroke events. Furthermore, a reduction in DBP was also associated with a lower risk of developing HF. Evident J-shaped curves were not observed in restricted cubic splines, suggesting that BP reduction did not cause significant harm in this cohort.
Most studies that assessed clinical outcomes associated with IDH in young adults have shown its association with adverse outcomes, including composite CVD events, atrial fibrillation, HF, and chronic kidney disease.4,6,13,14 Whereas, clinical outcomes associated with IDH in middle-aged and older adults were inconsistent. The discrepancy may be explained by differences in the mechanism underlying DBP levels between younger and older adults. For example, higher DBP in younger adults would reflect increased peripheral vascular resistance in young adults, whereas both arterial stiffening (i.e. lowered DBP) and peripheral vascular resistance (i.e. increased DBP) are reflected by DBP levels in older adults. In older adults, low DBP could also imply cardiac dysfunction.15 The Atherosclerosis Risk in Communities (ARIC) study mainly included middle-aged or older individuals (mean age, 55 years), and McEvoy et al.,1 reported no significant relationship between IDH and incident CVD. In another study using the ARIC dataset, DBP of < 60 mmHg was associated with higher high-sensitivity cardiac troponin-T and a higher risk of developing coronary heart disease.16 These results suggest the complicated clinical significance of DBP among middle-aged or older people. On the other hand, the present study mainly included relatively younger participants, and ‘reverse causation’ regarding the association between DBP and incident CVD was not evident.
The present study has some limitations and the potential study limitations are primarily attributed to the use of the JMDC Claims Database as we previously discussed.6,7 We used BP data measured at a health check-up (single occasion), and rigorous contemporary standard methods were not performed for measurements. Therefore, the BP data used in this study may not represent the BP phenotype of each individual, and the possibility of misclassification of BP status (e.g. categorization of IDH) cannot be eliminated. Due to the retrospective nature of the study, there is the possibility of unmeasured confounders and residual bias even after multivariable adjustment. For example, the JMDC Claims Database does not include data on socioeconomic status, which may have affected our results. Further, there were substantial differences in clinical characteristics between individuals with and without reduction in DBP, and therefore, these differences could influence the study results even after multivariable analyses. Given the age of the population in the JMDC Claims Database [the median age in the present study: 48 (42–54) years], the incidence rate of CVD in this dataset would be similar to other epidemiological data in Japan.17 Further, the accuracy (particularly, specificity) of recorded diagnoses in a Japanese administrative dataset was high.18 For example, a validation study on diagnosis in administrative claims in Japan showed that the positive predictive value of ICD-10 codes was ∼79% for MI, 89% for stroke, and 96% for HF.19 But, diagnoses recorded in insurance claims databases (such as the JMDC Claims Database) have not been fully verified. Additionally, it would be a concern that the JMDC Claims Database includes diagnoses made in outpatient as well as inpatient settings. Further investigations are needed to confirm our primary findings. The JMDC Claims Database mainly consists of employed individuals, and thus, the potential selection bias should be considered. Furthermore, the association between IDH and incident CVD was reportedly attenuated with age in particular.2,3 Although a reduction in DBP was associated with a decreased risk of CVD development even in people aged ≥50 years in the present study, the clinical benefit of BP reduction for older individuals with IDH should be validated using other independent datasets. In addition, our results indicated that a reduction in DBP was associated with a lower incidence of stroke or HF and was not associated with a risk of MI or all-cause death. The potential benefit of reduction in DBP among people having IDH varies depending on the targeted outcome. We could not analyse the association between changes in BP status since the index date (BP measurement at 1 year) and CVD risk. The serial change in BP status may have influenced the results. Moreover, a secondary analysis of SPRINT demonstrated that the initial beneficial effect of intensive treatment was lost over time.20 In addition to the serial change in BP, cumulative BP may have influenced the results. We could not define dyslipidaemia based on the 2021 European Society of Cardiology Guidelines.21 The diagnosis of dyslipidaemia may need to be personalized to individual clinical characteristics based on these clinical guidelines. Finally, data on CVD-related death were not available in the JMDC Claims Database.
Conclusions
The present study using a large-scale health check-up and administrative claims dataset revealed an association between a reduction in DBP and a reduced risk of CVD development in adults with IDH. We need to establish the optimal BP management strategy for individuals with IDH.
Author contributions
H.K., Y.Y., A.O., K.F., K.N., H.Y., and I.K. contributed to the conception and design of this work. Y.S., A.O., S.M., N.M., T.J., and H.Y. contributed to the analysis of the data for this work. H.K., Y.Y., A.O., K.F., N.T., H.M., K.N., H.Y., and I.K. contributed to the interpretation of the data. H.K., A.O., Y.S., N.T., H.M., and H.Y. drafted the manuscript. N.T., H.M., K.N., H.Y., and I.K. contributed to the critical revision for important intellectual content. H.Y. and I.K. approved the submission to this journal. All authors gave final approval and agreed to be accountable for all aspects of work, ensuring integrity and accuracy.
Supplementary material
Supplementary material is available at European Journal of Preventive Cardiology.
Funding
The Ministry of Education, Culture, Sports, Science and Technology of Japan (21AA2007) and the Ministry of Health, Labour and Welfare of Japan (21AA2007) provided grants in support of this work (20H03907, 21H03159, and 21K08123). The funder had no role in the study design, analysis, and interpretation of the data.
Data availability
The JMDC Claims Database is available for purchase from JMDC Inc. (https://www.eng.phm-jmdc.com/).
References
Author notes
Conflict of interest: Research grants and scholarships were provided to H.K. and K.F. by Boston Scientific Japan, Biotronik Japan, SIMPLEX QUANTUM, and Fukuda Denshi in Central Tokyo. A.O. is a member of the Department of Prevention of Diabetes and Lifestyle-related Diseases, which is a joint programme between The University of Tokyo and Asahi Mutual Life Insurance Company. There is nothing to disclose from the remaining authors. The funding sources had nothing with regard to the current study.
Comments