Blood Pressure and Cardiovascular Disease in Older Patients With Diabetes: Retrospective Cohort Study.

Background Blood pressure (BP) targets in elderly patients with diabetes remain unclear. We evaluated the association between BP and cardiovascular disease in elderly patients with diabetes without cardiovascular disease or heart failure. Methods and Results We performed a retrospective cohort study of 225 563 elderly (aged ≥65 years) patients with diabetes without cardiovascular disease or heart failure from 2009 to 2017 using the National Health Information Database. We divided the participants by systolic BP (SBP) and diastolic BP. Primary composite outcomes were stroke, myocardial infarction, heart failure, and all-cause death analyzed by Cox proportional hazards regression analysis adjusted for baseline covariates. During a median follow-up of 7.76 years, the incidence rate of primary composite outcomes was 26.62 per 1000 person-years. In multivariable Cox proportional hazard modeling, the risk of the primary outcome had a U-curved association with SBP/diastolic blood pressure with a nadir between 120 and 129 mm Hg/65 and 69 mm Hg, respectively. Hypertension medication was associated with lower risk of primary composite outcomes in SBP ≥140 mm Hg (P for interaction for SBP <0.001) and diastolic blood pressure ≥90 mm Hg (P for interaction for diastolic blood pressure=0.018). In participants aged ≥80 years, SBP ≥160 mm Hg was only a marginally higher risk for primary composite outcomes (hazard ratio=1.11; 95% CI, 0.98-1.24). Conclusions In this large sample of older Korean patients with diabetes, cardiovascular events were more common in people with resting SBP or diastolic BP ≥140 or 95 mm Hg, respectively, and also more common in people with resting SBP or diastolic BP <120 or 65 mm Hg, respectively.

Charlson Comorbidity Index

diastolic blood pressure

National Health Information Database

systolic blood pressure

Clinical Perspective

What Is New?

In this large sample of older Korean patients with diabetes, the risk of cardiovascular events had a U‐curved association with systolic blood pressure/diastolic blood pressure with a nadir between 120 and 129 mm Hg/65 and 69 mm Hg, respectively.

Hypertension medication was associated with lower risk of cardiovascular events in systolic blood pressure/diastolic blood pressure ≥140 mm Hg/ ≥90 mm Hg, respectively, and with higher risk in systolic blood pressure/diastolic blood pressure <110 mm Hg/<60 mm Hg, respectively.

What Are the Clinical Implications?

This study finding supports the recent guideline that recommends initiating hypertension medication in elderly patients with diabetes and systolic blood pressure/diastolic blood pressure ≥140 mm Hg/≥90 mm Hg, respectively.

Continuing antihypertensive therapy in elderly patients with diabetes and relative hypotension may increase cardiovascular events.

In patients with diabetes, hypertension is common (60%–85%) , , and is a major risk factor for both macrovascular and microvascular complications. , There is considerable evidence of the benefits of blood pressure (BP) control in patients with diabetes with hypertension to reduce the major macrovascular and microvascular complications of diabetes, as well as reduce mortality. , , A meta‐analysis of 13 randomized controlled trials involving patients with diabetes or prediabetes showed that a reduction in systolic blood pressure (SBP) to 131 to 135 mm Hg reduced the risk of all‐cause mortality by 13%, whereas more intensive SBP control (≤130 mm Hg) was only associated with a greater reduction in strokes.

The global population aged ≥60 years has more than doubled since 1980. One‐third of older people (aged >60 years) have diabetes. The association between elevated SBP and the risk of cardiovascular morbidity and mortality in the older population is also significantly positive. , Hence, BP control is also expected to offer benefits in older people and becomes increasingly important in the prevention of cardiovascular disease (CVD). However, little information is available on target BP levels in older patients with hypertension with type 2 diabetes, and there is some discrepancy in BP targets among guidelines. , , ,

In this study, we evaluated the potential relationships between BP level, CVD events, and mortality in older patients with diabetes without CVD using a large‐scale population data set from the National Health Information Database (NHID).

Methods

Anonymized data and materials have been made publicly available through National Health Insurance Sharing Service and can be accessed at https://nhiss.nhis.or.kr/.

Study Database

Data used in our analysis were from the NHID, a public database on health care use and health screening that contains sociodemographic and mortality information for the entire population of South Korea. The NHID contains data for the years 2002 to 2017. The NHID, which is produced by the National Health Insurance Service, was launched in 2000 by integrating 375 insurance associations. The National Health Insurance Service provides, through the national health screening program, regular health checkups and cancer screenings biannually at no cost for all insured Koreans aged >40 years. The NHID provides longitudinal data for 97% of the Korean population, with linkage to the National Death Registry and the national health screening program. , This latter program was initiated in 2009 and includes a medical interview and postural examination, chest x‐ray examination, blood test (including fasting plasma glucose and triglyceride levels), urine test, dental screening, and other tests. Approval for the study protocol was obtained from the institutional review board of Hanyang University Guri Hospital (GURI 2020‐08‐016). The need for informed consent was waived by the board.

Study Participants

This was a retrospective cohort study that included 249 903 individuals. From the NHID, 922 061 people with diabetes participated in the national health screening program in 2009. Among the 922 061 participants, 642 042 individuals aged <65 years, 12 250 patients with a history of stroke, 6843 patients with a history of myocardial infarction (MI), 24 340 patients with a history of heart failure (HF), and 11 023 individuals lacking complete data were excluded from our study. Therefore, the total number of eligible participants was 225 563. Participants were categorized into 8 groups by SBP and 9 groups by diastolic blood pressure (DBP).

Definitions of Diabetes and Study Outcomes (Cardiovascular Events and Death)

Patients with type 2 diabetes were identified from the insurance claims data as having at least 1 claim per year for the prescription of antidiabetic medication under International Classification of Diseases, Tenth Revision (ICD‐10) diagnostic codes E11 to E14, or from national health screening program data as having fasting plasma glucose of ≥126 mg/dL. The primary outcome of the study was a composite of nonfatal stroke, nonfatal MI, nonfatal HF, and all‐cause death. Secondary outcomes were newly diagnosed MI, stroke, HF, or all‐cause death. Stroke was defined as an ICD‐10 code I63 or I64 during hospitalization with claims for brain magnetic resonance imaging or brain computerized tomography. Nonfatal MI was defined as an ICD‐10 code I21 or I22 during hospitalization. Nonfatal HF was defined as an ICD‐10 code I50 during hospitalization. The study population was followed from baseline to the date of death, onset of a cardiovascular event, or until December 31, 2017, whichever came first.

Clinical and Laboratory Measurements

All participants completed a questionnaire on medical history, use of tobacco and alcohol, and exercise habits. Smoking habits were categorized as nonsmoker, ex‐smoker, or current smoker. Alcohol consumption was classified as nondrinker, moderate drinker (<30 g per day), or heavy drinker (≥30 g per day). Regular exercise was defined as vigorous exercise 3 or more times per week or moderate exercise 5 or more times per week. Body mass index (BMI) was calculated as body weight (in kilograms) divided by height (in meters squared). BP was measured through a standard national health screening program protocol. All participants had rested for at least 5 minutes in a seated position before the first measurement. If SBP was >120 mm Hg or DBP was >80 mm Hg, remeasurement was performed after an interval of 2 minutes or more. BP was measured by the auscultation method using a stethoscope or using an oscilloscopic automatic sphygmomanometer in a quiet environment. The device was calibrated daily. A cuff with an appropriately sized bladder was used. The standard bladder for adults is 12 cm wide and 26 cm long. A bladder with a width of at least 40% of the circumference of the arm and a length of 80% to 100% of the circumference of the arm was used. Examinees were recommended to avoid smoking, alcohol, or caffeine before measurement. Blood samples were collected after overnight fasting. Plasma glucose, total cholesterol, triglycerides, high‐density lipoprotein cholesterol, and low‐density lipoprotein cholesterol were measured. We calculated glomerular filtration rate using the 4‐variable Modification of Diet in Renal Disease Study equation. Baseline comorbidities were identified as dyslipidemia (ICD‐10 code E78 with lipid‐lowering agents or serum total cholesterol ≥240 mg/dL), and chronic kidney disease (CKD) (estimated as a glomerular filtration rate <60 mL/min per 1.73 m2). The Charlson Comorbidity Index (CCI) was used to estimate the comorbidity burden by reviewing the ICD‐10 codes. An individual was considered to have a comorbidity when the respective diagnostic codes were present >2 times within 1 year before the inclusion date.

Statistical Analysis

Baseline characteristics were analyzed using descriptive statistics. Categorical variables were described as frequency and percentage. Continuous variables were described as mean±SD for normally distributed data and as the geometric mean and 95% CI for data not normally distributed. We compared the baseline characteristics of 8 groups categorized by SBP and 9 groups categorized by DBP. Continuous variables were compared using 1‐way ANOVA, whereas categorical variables were compared using the Mantel‐Haenszel χ2 test. The follow‐up duration of each group was obtained. The incidence rates for stroke, MI, HF, and death were estimated for each group over the total follow‐up period. Incidence curves were estimated using the Kaplan‐Meier method, and the log‐rank test was also conducted. All outcomes were analyzed by Cox proportional hazards regression analysis while controlling for baseline covariates. The proportional hazard assumption was assessed by visual inspection of the scaled Schoenfeld residuals plot and log‐log survival plot. We deemed a 2‐tailed P value <0.05 to be significant. Analyses were performed with SAS 9.4 (SAS Institute, Cary, NC) and R version 3.4.1 (The R Foundation for Statistical Computing, Vienna, Austria; http://www.R‐project.org).

Results

Baseline Characteristics of Participants

The characteristics of the 8 SBP groups are described in Table 1. The group with higher SBP were older; more likely to be women, heavy drinkers, and hypertension medication users; had a higher BMI and fasting plasma glucose level; and had a higher prevalence of dyslipidemia and CKD. The group with higher SBP was less likely to comprise current smokers, regular exercisers, and insulin users. Similar patterns of baseline characteristics were noted in the 9 DBP groups (Table S1). Over half of the patients with SBP <120 mm Hg (55.6%) or DBP <65 mm Hg (59.6%) were prescribed hypertension medication.

Table 1

Baseline Demographic and Clinical Characteristics in SBP Groups

Characteristic	SBP, mm Hg
	<100	100–110	110–120	120–130	130–140	140–150	150–160	>160	P value
No. of subjects	2228	9716	36 295	51 706	78 308	32 931	21 363	17 356
Sex									<0.0001
Men	1109 (49.78)	4723 (48.61)	17 689 (48.74)	24 958 (48.27)	37 909 (48.41)	15 861 (48.16)	10 338 (48.39)	7975 (45.95)
Women	1119 (50.22)	4993 (51.39)	18 606 (51.26)	26 748 (51.73)	40 399 (51.59)	17 070 (51.84)	11 025 (51.61)	9381 (54.05)
Smoker									<0.0001
Non	1467 (65.84)	6574 (67.66)	25 105 (69.17)	36 446 (70.49)	55 937 (71.43)	23 549 (71.51)	15 505 (72.58)	12 911 (74.39)
Ex	364 (16.34)	1548 (15.93)	5802 (15.99)	8499 (16.44)	12 650 (16.15)	5482 (16.65)	3468 (16.23)	2497 (14.39)
Current	397 (17.82)	1594 (16.41)	5388 (14.85)	6761 (13.08)	9721 (12.41)	3900 (11.84)	2390 (11.19)	1948 (11.22)
Alcohol consumption									<0.0001
Non	1786 (80.16)	7693 (79.18)	27 900 (76.87)	39 116 (75.65)	58 172 (74.29)	24 037 (72.99)	15 364 (71.92)	12 614 (72.68)
Mild	364 (16.34)	1773 (18.25)	7142 (19.68)	10 664 (20.62)	16 786 (21.44)	7331 (22.26)	4859 (22.74)	3774 (21.74)
Heavy	78 (3.5)	250 (2.57)	1253 (3.45)	1926 (3.72)	3350 (4.28)	1563 (4.75)	1140 (5.34)	968 (5.58)
Regular exercise	815 (36.58)	3795 (39.06)	14 472 (39.87)	20 582 (39.81)	30 858 (39.41)	13 088 (39.74)	7965 (37.28)	6027 (34.73)	<0.0001
Dyslipidemia	879 (39.45)	4078 (41.97)	14 818 (40.83)	21 886 (42.33)	32 938 (42.06)	14 134 (42.92)	9119 (42.69)	7462 (42.99)	<0.0001
Chronic kidney disease	616 (27.65)	2328 (23.96)	8267 (22.78)	11 582 (22.4)	17 610 (22.49)	7557 (22.95)	4828 (22.6)	4243 (24.45)	<0.0001
Insulin	314 (18.62)	1315 (17.44)	4256 (15.4)	5620 (14.33)	8033 (13.7)	3345 (13.85)	2102 (13.87)	1779 (14.82)	<0.0001
No. of oral diabetes medications									<0.0001
0	47 (2.79)	183 (2.43)	577 (2.09)	744 (1.9)	1167 (1.99)	544 (2.25)	305 (2.01)	309 (2.57)
1	498 (29.54)	2119 (28.1)	7958 (28.79)	11 739 (29.94)	17 903 (30.52)	7498 (31.05)	4675 (30.86)	3652 (30.43)
2	644 (38.2)	3091 (40.99)	11 453 (41.44)	16 454 (41.96)	24 916 (42.48)	10 315 (42.72)	6603 (43.58)	5200 (43.33)
3	403 (23.9)	1737 (23.04)	6195 (22.41)	8446 (21.54)	12 230 (20.85)	4800 (19.88)	2989 (19.73)	2412 (20.1)
4	80 (4.74)	353 (4.68)	1295 (4.69)	1647 (4.2)	2196 (3.74)	889 (3.68)	531 (3.5)	391 (3.26)
5	14 (0.83)	54 (0.72)	154 (0.56)	175 (0.45)	235 (0.4)	95 (0.39)	45 (0.3)	37 (0.31)
6	0 (0)	3 (0.04)	6 (0.02)	9 (0.02)	9 (0.02)	4 (0.02)	2 (0.01)	1 (0.01)
Hypertension medication	1130 (50.72)	5286 (54.41)	21 517 (59.28)	34 148 (66.04)	56 295 (71.89)	25 673 (77.96)	17 293 (80.95)	14 470 (83.37)	<0.0001
Mean±SD
Age, y	71.15±5.08	70.88±4.79	71±4.81	71.02±4.78	71.14±4.77	71.19±4.83	71.27±4.83	71.59±5.01	<0.0001
BMI, kg/m2	22.86±3.24	23.48±3.16	24.08±3.11	24.44±3.1	24.76±3.13	24.98±3.16	25.01±3.2	25±3.33	<0.0001
FBG, mg/dL	134.56±44.71	134.7±43.31	135.05±42.49	134.03±40.86	134.83±40.48	135.98±40.39	137.02±40.99	139.31±42.65	<0.0001
SBP, mm Hg	92.5±4.33	103±3.32	113.52±3.64	122.55±3.17	132.97±3.42	142.12±2.9	151.72±2.72	167.3±9.08	<0.0001
DBP, mm Hg	59.44±5.74	64.03±6.29	70.11±6.35	74.66±6.96	78.8±7.11	82.82±8.35	86.75±8.99	92.08±10.24	<0.0001
Median, Q1–Q3
Age, y	70 (67–74)	70 (67–74)	70 (68–74)	70 (68–74)	70 (68–74)	70 (68–74)	70 (68–74)	70 (68–74)	<0.0001
BMI, kg/m2	22.77 (20.7–24.91)	23.42 (21.34–25.46)	24 (22.03–25.97)	24.3 (22.38–26.33)	24.61 (22.68–26.64)	24.78 (22.89–26.84)	24.86 (22.89–26.93)	24.84 (22.83–27.01)	<0.0001
FBG, mg/dL	128 (105–149)	128 (107–149)	128 (108–149)	128 (108–148)	129 (109–149)	130 (110–150)	131 (111–151)	132 (113–154)	<0.0001
SBP, mm Hg	91 (90–96)	101.5 (100–106)	112 (110–118)	120 (120–125)	131 (130–136)	140 (140–144)	150 (150–153)	164 (160–170)	<0.0001
DBP, mm Hg	60 (56–61)	62 (60–70)	70 (67–74)	76 (70–80)	80 (74–82)	81 (79–90)	90 (80–90)	90 (86–100)	<0.0001

BMI indicates body mass index; DBP, diastolic blood pressure; FBG, fasting blood glucose; and SBP, systolic blood pressure.

Blood Pressure and Primary Composite Outcomes

There were 77 447 (34.3%) primary composite outcomes in the 7.76‐year mean follow‐up period (Table 2). The incidence rate of primary composite outcomes was 47.90 per 1000 person‐years. In patients with SBP 120 to 129 mm Hg, the incidence rate of primary composite outcomes was 45.2 per 1000 person‐years and lower than in the other SBP groups (Table 2 and Figure S1). In multivariable Cox proportional hazard models of the SBP reference group (SBP 120–129 mm Hg), the risk of the primary outcome increased significantly with not only higher SBP but also lower SBP after adjusting for age, sex, smoking, alcohol consumption, regular exercise, BMI, dyslipidemia, CKD, insulin treatment, number of oral diabetes medications, fasting plasma glucose level, CCI, hypertension medication, and DBP (Figure 1A). In patients with DBP 65 to 69 mm Hg, the incidence rate of primary composite outcomes was 43.2 per 1000 person‐years and lower than in the other DBP groups (Table 2 and Figure S1). The risk of the primary outcome increased significantly in not only higher but also lower DBP groups than the DBP reference group (DBP 65–69 mm Hg) in multivariable Cox proportional hazard modeling with age, sex, smoking, alcohol consumption, regular exercise, BMI, dyslipidemia, CKD, insulin treatment, number of oral diabetes medications, fasting plasma glucose level, CCI, hypertension medication, and SBP (Figure 1B).

Table 2

Number, Incidence Rate, and HR for Primary Composite Outcomes (Myocardial Infarction, Stroke, and All‐Cause Mortality) Stratified by SBP and DBP

	No. of patients	No. of events	Duration, person‐years	Rate, events per 1000 person‐years	HR (95% CI)	No. of patients	No. of events	Duration, person‐years	Rate, events per 1000 person‐years
SBP, mm Hg
<100	1932	805	12 892	62.4	1.36 (1.25–1.47)	1.20 (1.11–1.30)	1.20 (1.10–1.30)	1.22 (1.12–1.33)	1.29 (1.12–1.40)
100–109	8612	3026	61 054	49.6	1.11 (1.07–1.16)	1.04 (0.99–1.08)	1.03 (0.98–1.08)	1.04 (1.00–1.09)	1.09 (1.04–1.14)
110–119	32 551	11 059	233 087	47.4	1.05 (1.02–1.08)	1.03 (1.00–1.06)	1.02 (0.99–1.05)	1.03 (1.00–1.06)	1.05 (1.02–1.08)
120–129	46 530	15 226	337 133	45.2	1 (Ref.)	1 (Ref.)	1 (Ref.)	1 (Ref.)	1 (Ref.)
130–139	70 892	23 815	510 965	46.6	1.04 (1.02–1.06)	1.06 (1.03–1.08)	1.06 (1.03–1.08)	1.05 (1.02–1.07)	1.03 (1.01–1.06)
140–149	30 001	10 339	215 231	48.0	1.08 (1.04–1.11)	1.11 (1.08–1.14)	1.11 (1.07–1.14)	1.09 (1.06–1.12)	1.06 (1.03–1.09)
150–159	19 351	6921	137 772	50.2	1.13 (1.09–1.16)	1.16 (1.12–1.20)	1.16 (1.12–1.20)	1.14 (1.10–1.17)	1.09 (1.05–1.13)
≥160	15 694	6256	108 739	57.5	1.31 (1.27–1.36)	1.35 (1.30–1.39)	1.33 (1.29–1.38)	1.30 (1.27–1.35)	1.22 (1.17–1.27)
DBP, mm Hg
<60	3486	1332	24 308	54.8	1.21 (1.14–1.30)	1.12 (1.05–1.20)	1.11 (1.04–1.19)	1.12 (1.04–1.19)	1.14 (1.07–1.22)
60–64	15 631	5606	110 546	50.7	1.14 (1.09–1.19)	1.10 (1.05–1.15)	1.09 (1.04–1.14)	1.09 (1.05–1.14)	1.11 (1.06–1.16)
65–69	14 266	4481	103 684	43.2	1 (Ref.)	1 (Ref.)	1 (Ref.)	1 (Ref.)	1 (Ref.)
70–74	48 346	16 459	346 971	47.4	1.10 (1.06–1.14)	1.10 (1.06–1.14)	1.10 (1.06–1.14)	1.10 (1.06–1.14)	1.10 (1.06–1.14)
75–79	24 501	7824	177 759	44.0	1.05 (1.01–1.09)	1.08 (1.03–1.12)	1.08 (1.04–1.13)	1.08 (1.03–1.12)	1.07 (1.02–1.11)
80–84	64 746	22 307	464 584	48.0	1.13 (1.09–1.17)	1.15 (1.11–1.19)	1.15 (1.11–1.19)	1.14 (1.10–1.19)	1.13 (1.09–1.17)
85–89	18 792	6313	135 544	46.6	1.13 (1.08–1.18)	1.17 (1.12–1.23)	1.18 (1.13–1.23)	1.16 (1.11–1.22)	1.14 (1.09–1.19)
90–94	23 623	8490	168 341	50.4	1.21 (1.16–1.26)	1.25 (1.20–1.31)	1.25 (1.20–1.31)	1.23 (1.18–1.29)	1.19 (1.14–1.25)
≥95	12 172	4635	85 137	54.4	1.36 (1.29–1.42)	1.41 (1.34–1.47)	1.40 (1.34–1.47)	1.37 (1.31–1.44)	1.31 (1.24–1.38)

Model 1: age, and sex. Model 2: age, sex, smoking, alcohol consumption, regular exercise, body mass index, dyslipidemia, chronic kidney disease, and Charlson Comorbidity Index. Model 3: age, sex, smoking, alcohol consumption, regular exercise, body mass index, dyslipidemia, chronic kidney disease, Charlson Comorbidity Index, insulin treatment, number of oral diabetes medications, and fasting plasma glucose. Model 4: age, sex, smoking, alcohol consumption, regular exercise, body mass index, dyslipidemia, chronic kidney disease, Charlson Comorbidity Index, insulin treatment, number of oral diabetes medications, fasting plasma glucose, and hypertension medication. Model 5: age, sex, smoking, alcohol consumption, regular exercise, body mass index, dyslipidemia, chronic kidney disease, Charlson Comorbidity Index, insulin treatment, number of oral diabetes medications, fasting plasma glucose, hypertension medication, and systolic or diastolic blood pressure. DBP indicates diastolic blood pressure; HR, hazard ratio; Ref., reference; and SBP, systolic blood pressure.

Figure 1

Hazard ratio (HR) and 95% CI for primary composite outcomes by (A) systolic blood pressure and (B) diastolic blood pressure.

Adjusted for age, sex, smoking, alcohol consumption, regular exercise, body mass index, dyslipidemia, chronic kidney disease, insulin treatment, number of oral diabetes medications, fasting plasma glucose level, Charlson Comorbidity Index, hypertension medication, and systolic or diastolic blood pressure.

Blood Pressure and Myocardial Infarction, Stroke, Heart Failure, and All‐Cause Death

There were 11 683 (5.18%) incidents of MI in the 7.61‐year mean follow‐up period (Table S2). The incidence rate of MI was 6.80 per 1000 person‐years. In patients with SBP 120 to 129 mm Hg, the incidence rate of MI was 6.39 per 1000 person‐years and lower than in the other SBP groups (Table S2 and Figure S2A). The incidence rate of MI in patients with DBP 75 to 79 mm Hg was 6.01 per 1000 person‐years and lower than in the other DBP groups (Table S2 and Figure S2A). In multivariable Cox proportional hazard modeling, the risk of MI was higher in patients with type 2 diabetes and SBP ≥140 mm Hg (Figure 2A) or DBP ≥90 mm Hg (Figure 2B) compared with the reference group (SBP 120–129 mm Hg, DBP 65–69 mm Hg) after adjusting for age, sex, smoking, alcohol consumption, regular exercise, BMI, dyslipidemia, CKD, insulin treatment, number of oral diabetes medications, fasting plasma glucose level, CCI, hypertension medication, and SBP or DBP.

Figure 2

Hazard ratio (HR) and 95% CI for myocardial infarction by (A) systolic blood pressure and (B) diastolic blood pressure, stroke by (C) systolic blood pressure and (D) diastolic blood pressure, heart failure by (E) systolic blood pressure and (F) diastolic blood pressure, and all‐cause death by (G) systolic blood pressure and (H) diastolic blood pressure.

Adjusted for age, sex, smoking, alcohol consumption, regular exercise, body mass index, dyslipidemia, chronic kidney disease, insulin treatment, number of oral diabetes medications, fasting plasma glucose level, Charlson Comorbidity Index, hypertension medication, and systolic or diastolic blood pressure.

There were 22 572 (10.00%) incidents of stroke in the 7.45‐year mean follow‐up period (Table S2). The incidence rate of stroke was 13.44 per 1000 person‐years. In patients with SBP 120 to 129 mm Hg, the incidence rate of stroke was 12.06 per 1000 person‐years and lower than in the other SBP groups (Table S2 and Figure S2B). The incidence rate of stroke in patients with DBP 65 to 69 mm Hg was 11.31 per 1000 person‐years and lower than in the other DBP groups (Table S2 and Figure S2B). In multivariable Cox proportional hazard modeling, the risk of stroke was increased with higher BP and lower BP (SBP <100 mm Hg, and DBP <65 mm Hg) compared with the reference group (SBP 120–129 mm Hg, DBP 65–69 mm Hg) after adjusting for age, sex, smoking, alcohol consumption, regular exercise, BMI, dyslipidemia, CKD, insulin treatment, number of oral diabetes medications, fasting glucose level, CCI, hypertension medication, and SBP or DBP (Figure 2C and 2D).

There were 24 190 (10.72%) incidents of HF in the 7.55‐year mean follow‐up period (Table S2). The incidence rate for HF was 14.19 per 1000 person‐years. In patients with SBP 120 to 129 mm Hg, the incidence rate of HF was 13.40 per 1000 person‐years and lower than in the other SBP groups (Table S2 and Figure S2C). The incidence rate of HF in patients with DBP 75 to 79 mm Hg was 12.66 per 1000 person‐years and lower than in the other DBP groups (Table S2 and Figure S2C). In multivariable Cox proportional hazard modeling, the risk of HF was higher in patients with type 2 diabetes and SBP ≥150 mm Hg or SBP <100 mm Hg (Figure 2E) and DBP ≥95 mm Hg or DBP <60 mm Hg (Figure 2F) compared with the reference group (SBP 120–129 mm Hg, DBP 65–69 mm Hg) after adjusting for age, sex, smoking, alcohol consumption, regular exercise, BMI, dyslipidemia, CKD, insulin treatment, number of oral diabetes medications, fasting plasma glucose level, CCI, hypertension medication, and SBP or DBP.

There were 46 626 (20.67%) incidents of all‐cause death in the 7.76‐year mean follow‐up period (Table S2 and Figure S2G). The incidence rate for all‐cause death was 26.6 per 1000 person‐years. In patients with SBP 120–129 mm Hg, the incidence rate of all‐cause death was 25.22 per 1000 person‐years and lower than in the other SBP groups (Table S2 and Figure S2G). The incidence rate of stroke in patients with DBP 65–69 mm Hg was 23.68 per 1000 person‐years and lower than in the other DBP groups (Table S2 and Figure S2B). In multivariable Cox proportional hazard modeling, the risk of all‐cause death was increased in older patients with type 2 diabetes and SBP ≥140 mm Hg or <120 mm Hg (Figure 2G), and DBP ≥70 mm Hg or <65 mm Hg (Figure 2H) compared with the reference group (SBP 120–129 mm Hg, DBP 65–69 mm Hg) after adjusting for age, sex, smoking, alcohol consumption, regular exercise, BMI, dyslipidemia, CKD, insulin treatment, number of oral diabetes medications, fasting plasma glucose level, CCI, hypertension medication and SBP or DBP.

Sensitivity and Subgroup Analysis

Sensitivity analysis after excluding the first year of observation had consistent findings: the risk of primary composite outcomes was lowest in patients with SBP 120 to 129 mm Hg and with DBP 65 to 69 mm Hg, and the risk of primary composite outcomes increased significantly with not only higher but also lower SBP and DBP than the reference group, with SBP 120 to 129 mm Hg or DBP 65 to 69 mm Hg (Figure 3A and 3B).

Figure 3

Hazard ratio (HR) and 95% CI for primary composite outcomes by (A) systolic blood pressure and (B) diastolic blood pressure after excluding the first year of observation, by (C) systolic blood pressure and (D) diastolic blood pressure in subgroup with aged ≥80 years or with aged <80 years, and by (E) systolic blood pressure and (F) diastolic blood pressure in subgroup with or without hypertension medication.

Adjusted for age, sex, smoking, alcohol consumption, regular exercise, body mass index, dyslipidemia, chronic kidney disease, insulin treatment, number of oral diabetes medications, fasting plasma glucose level, Charlson Comorbidity Index, hypertension medication, and systolic or diastolic blood pressure.

Subgroup analysis of those aged ≥80 years and those aged <80 years revealed a significant interaction for the risk of the primary outcome according to SBP (P for interaction=0.014; Figure 3C), but not according to DBP (P for interaction=0.225; Figure 3D). The risk for primary composite outcomes was only marginally higher in participants aged ≥80 years and SBP ≥160 mm Hg (hazard ratio [HR], 1.11; 95% CI, 0.98–1.24; Figure 3C), but was significantly higher in participants aged <80 years than the reference group, with SBP 120 to 129 mm Hg (Figure 3C). Subgroup analyses were performed with hypertension medication and without hypertension medication for the risk of primary composite outcomes (Figure 3E and 3F). There was a significant interaction between the subgroups with or without hypertension medication for the risk of the primary outcome according to the SBP group (P for interaction <0.001; Figure 3E) and the DBP group (P for interaction=0.018; Figure 3F). Hypertension medication was associated with a lower risk for primary composite outcomes in SBP ≥140 mm Hg and DBP ≥90 mm Hg and higher risk for primary composite outcomes in SBP <110 mm Hg or DBP <60 mm Hg (Figure 3E and 3F). The spline curve for primary composite outcomes by SBP and DBP and with or without hypertension medication had consistent findings. The risk of secondary outcomes after excluding the first year of observation also had consistent findings (Figure 4).

Figure 4

Spline curve for composite primary outcomes by systolic or diastolic blood pressure and by systolic or diastolic blood pressure with or without hypertension medication adjusted for age, sex, smoking, alcohol consumption, regular exercise, body mass index, dyslipidemia, chronic kidney disease, insulin treatment, number of oral diabetes medications, fasting plasma glucose, hypertension medication, Charlson Comorbidity Index, and systolic or diastolic blood pressure.

HR indicates hazard ratio; and Ref, reference.

Discussion

In this large retrospective study, we examined the association between SBP and DBP and the risk of MI, stroke, HF and death in older patients with diabetes and without CVD. SBP 120 to 129 mm Hg and DBP 65 to 69 mm Hg presented the lowest risk of primary composite outcomes in all participants. Hypertension medication was associated with lower risk of primary composite outcomes in SBP ≥140 mm Hg and DBP ≥90 mm Hg and higher risk for primary composite outcomes in SBP <110 mm Hg and DBP <60 mm Hg. In participants aged ≥80 years, SBP ≥160 mm Hg was only marginally higher risk for primary composite outcomes.

Hypertension is common in older patients with type 2 diabetes. Although tight BP control is particularly beneficial in younger patients with diabetes, there is little evidence for BP goals in older patients with hypertension with type 2 diabetes. The ACCORD (Action to Control Cardiovascular Risk in Diabetes) BP trial has shown that intensive lowering of SBP (<120 mm Hg) in patients with hypertension with diabetes did not improve overall CVD events or deaths, except for a significant reduction in the risk for stroke compared with the conventional lowering of SBP (<140 mm Hg). In the same study, blood pressure control after the first year of follow‐up was 119.3/64.4 mm Hg in the intensive group and 133.5/70.5 mm Hg in the standard group. Other studies reported that further SBP lowering (up to 130–135 mm Hg) was associated with the lowest cardiovascular event rate. , , These findings were compatible with our study and were explained by a U‐curved association between SBP/DBP and primary composite outcomes in this study. The risk of primary composite outcomes in SBP 110 to 119 mm Hg (HR, 1.05; 95% CI, 1.02–1.08) and DBP 60 to 64 mm Hg (HR, 1.11; 95% CI, 1.06–1.16) was similar to that in SBP 130 to 139 mm Hg (HR, 1.03; 95% CI, 1.01–1.06) and DBP 70 to 74 mm Hg (HR, 1.10; 95% CI, 1.06–1.14) compared with the SBP and DBP reference groups. In addition, the risk of stroke in participants with SBP 130 to 139 mm Hg was significantly higher than the reference group (SBP 120–129 mm Hg), but the risk of MI, HF, and all‐cause death was not.

We also observed that SBP <120 mm Hg/DBP <65 mm Hg was associated with poor prognosis in patients with diabetes aged ≥65 years. A few studies in older patients with diabetes showed similar results. Post hoc analysis of the cohort of participants with diabetes in the INVEST (International Verapamil SR‐Trandolapril Study) (mean age >65 years) reported that SBP <110 mm Hg was associated with significantly increased risk (HR, 2.18; 95% CI, 1.17–4.09) of all‐cause mortality compared with SBP 125 to 130 mm Hg. A previous observational study of 1294 patients with type 2 diabetes aged ≥69.1 years showed a U‐shaped association between SBP and mortality, but not in those aged <69.1 years. Although hypertension is an established risk factor for HF, one study reported that low SBP (SBP <120 mm Hg) was associated with increased risks of HF among patients with diabetes without a history of HF. The HOT (Hypertension Optimal Treatment) study reported that lowering DBP to a target level of ≤80 mm Hg in patients with type 2 diabetes resulted in a 51% reduction in major CVD events compared with the conventional target (≤90 mm Hg). Other studies with isolated systolic hypertension, which is common in older people, showed a U‐curved association between DBP and CVD events, with a nadir between 60 and 70 mm Hg. , A recent cohort study of 1.3 million adults in a general outpatient population revealed a U‐curved relationship between the diastolic blood pressure and the composite outcome. The risk of composite outcomes was increased below 60 mm Hg, and the author suggested that this relationship was explained by age and other covariates. Our study results were also compatible with the above study. The association between DBP and the risk of the composite outcomes was U‐curved, with a nadir between 65 and 69 mm Hg. The risk of the composite outcomes was increased at <65 mm Hg. The association between lower blood pressure and poor outcomes could be explained by comorbidities common in older patients with lower blood pressure. Therefore, we adjusted for CCI in multivariable Cox proportional hazard modeling and performed sensitivity analyses excluding participants who were diagnosed with stroke, MI, HF, or all‐cause death during the first year of follow‐up. However, the association of lower blood pressure with increased risk of stroke, MI, HF, and all‐cause death in study participants did not change.

Two decades ago, the Framingham data suggested an age‐dependent threshold for hypertension. This study showed a different association between SBP and primary composite outcomes by age group (P for interaction=0.014). In patients with diabetes who were aged ≥80 years, the risk for primary composite outcomes was marginally higher in SBP ≥160 mm Hg than SBP 120 to 129 mm Hg (HR, 1.11; 95% CI, 0.98–1.24; Figure 3C). Based on our study results, the suggested SBP threshold is ≈130 mm Hg in patients with diabetes aged 65 to 79 years, and ≈160 mm Hg in patients with diabetes aged ≥80 years. Most guidelines do not indicate a BP goal in older patients with both diabetes and hypertension, because most hypertension trials do not include an older population, do not present age‐specific results, or have only a small number of older patients with diabetes in the study population. The American Diabetes Association recommends a treatment goal of <140/90 mm Hg in patients with diabetes (<130/90 mm Hg in patients with diabetes at higher cardiovascular risk), and cautiously suggests the same treatment goal in older patients with diabetes. , Interestingly, in subgroup analysis, hypertension medication was associated with lower risk of primary composite outcomes in SBP ≥140 mm Hg and DBP ≥90 mm Hg (Figure 3E and 3F). This finding supports the current American Diabetes Association cutoff value for initiating antihypertension treatment in elderly patients with diabetes. However, another cutoff is needed for patients with diabetes aged ≥80 years (Figure 3C).

Another concerning finding of this study is that the prevalence of prescribed hypertensive medications was quite high in patients with SBP <120 mm Hg (55.6%) and DBP <65 mm Hg (59.6%). Hypertension medication was associated with poor prognosis in patients with SBP <110 mm Hg or DBP <60 mm Hg (Figure 3E and 3F). One of 8 patients (12.9%) in our study who were on hypertensive medications might have been overtreated for hypertension. This finding suggests that continuing antihypertensive therapy in people with relative hypotension may increase, rather than decrease, cardiovascular events.

The strengths of our study are that we used a large‐scale nationwide database representing the entire Korean population. Second, we conducted fully adjusted analyses with all available confounding cardiovascular risk factors. However, this study also has some limitations. First, the retrospective observational study design has inherent limitations. Although the analyses were adjusted for most available demographic and clinical variables, some unidentified parameters could affect the results. Second, there was some possibility of selection bias, because we selected participants whose BP was measured in the national health screening program, which requests that participants voluntarily visit clinics. Therefore, there was some possibility that study participants might have higher mobility than individuals who did not participate in the national health screening program and, particularly in volunteers aged >75 years, may have less overall morbidity because of fewer serious chronic health conditions compared with the general elderly Korean population. However, our study excluded patients with stroke (including those with cerebral palsy) before analysis and might cover most elderly patients with diabetes in an outpatient setting. Third, we defined MI and stroke based on claims data; this may not be a completely accurate method for determining the number of cases. To overcome this problem, we defined outcomes using an operational definition by combining diagnostic and prescription records. Fourth, the BP measurement protocol of the national health screening program, which focused on screening hypertension, does not follow published guidelines and may have introduced measurement bias, which is likely to have estimated BP with reasonable accuracy in low or normal BP participants, but truncated the actual BP of individuals with high values. However, a recent study evaluated the performance of hypertension screening in medical institutions conducting the national health screening program in the Republic of Korea. According to this study, medical institutions used mainly oscilloscopic devices for BP measurement, and most had measurement manuals and training protocols. The majority of the institutions measured BP multiple times with a resting period and used the average values as an individual’s BP level. Fifth, this study is not a prospective study; therefore, causality cannot be determined. However, to minimize the possible effects of reverse causality, participants with preexisting MI or stroke were excluded. Lastly, we did not analyze the cause of death, because the details were unavailable in the national database.

In conclusion, this retrospective cohort study of older patients with diabetes without CVD suggests that patients with SBP 120 to 129 mm Hg and DBP 65 to 69 mm Hg had the lowest risk of primary composite outcomes for MI, stroke, HF, and all‐cause death. This study also suggested that hypertension medication may decrease CVD events in patients with SBP ≥140 mm Hg and DBP >90 mm Hg and increase CVD events in patients with SBP <110 mm Hg and DBP <60 mm Hg.

Sources of Funding

None.

Disclosures

None.

Tables S1–S2

Figures S1–S2

Click here for additional data file.