Heart disease (HD), cerebrovascular disease (CBD), and kidney disease (KD) are serious diseases. Ischemic HD (IHD), stroke (a type of CBD), and KD were the leading, second-leading, and tenth-leading causes of deaths globally in 2019, respectively (1). The World Health Organization (WHO) estimated (1) that IHD caused 8.9 million or 16% of the world's total deaths, stroke caused 6.2 million deaths, approximately 11% of the total deaths, and the deaths caused by KD totaled 1.3 million in 2019.

In the United States, HD, CBD (stroke), and KD were the leading, fifth leading, and tenth-leading causes of death in 2020. They caused 696,962 (20.6%), 160,264 (4.7%), and 52,547 (1.6%) deaths, respectively (percentages of total deaths are listed in parentheses) (2).

In Japan, HD, CBD, and KD were the second-, fourth-, and eighth-leading causes of death in 2020 (3). They caused 205,596 (15.0%), 102,978 (7.5%), and 26,948 (2.0%) deaths in 2020, respectively (percentages of total deaths are listed in parentheses). The medical expenditures in fiscal year 2019 (4) were 2.09 trillion yen for HD, 1.83 trillion yen for CBD, and 1.66 trillion yen for KD, respectively. These medical expenditures accounted for 12.6% of the total Japanese national medical expenditure of 44.39 trillion yen.

The American Heart Association (AHA) (5) described the risk factors for coronary HD. The major risk factors are classified into three categories: (i) non-modifiable risk factors that cannot be changed; (ii) modifiable risk factors that can be modified, treated, or controlled; and (iii) other factors that contribute to HD risks. Non-modifiable risk factors include age, sex, and heredity (including race). Modifiable risk factors include tobacco smoking, high blood cholesterol, total cholesterol, triglycerides, high blood pressure or hypertension, physical inactivity, obesity, being overweight, and diabetes. Other factors included stress, alcohol consumption, diet, and nutrition. Centers for Disease Control and Prevention (CDC) (6) described that the risk factors for HD and stroke are high blood pressure, low-density lipoprotein (LDL) cholesterol, diabetes, smoking and secondhand smoke exposure, obesity, unhealthy diet, and physical inactivity.

High blood pressure (BP) or hypertension is considered a major risk factor for HD or cardiovascular disease (CVD). Fuchs and Whelton (7) noted that hypertension had the strongest evidence of causation among the risk factors for CVD. The American College of Cardiology (ACC), AHA, and other organizations (8) presented a new guideline (the 2017ACC/AHA guideline) for hypertension in 2017. Under the new guideline, the criteria for hypertension lowered to 130/80 mmHg (SBP ≥ 130 or DBP ≥ 80 mmHg) from 140/90 mmHg (SBP ≥140 or DBP ≥ 90 mmHg) of the JNC7 guideline settled in 2003 (9). Nawata (10) reported that 14.0% and 38.0% were classified as having hypertension according to the 140/90 and 130/90 mmHg criteria, respectively. Muntner et al. (11) mentioned that the 2017 ACC/AHA guideline would increase medication use and reduce the prevalence of CVD.

Many studies have analyzed the relationship between high BP or hypertension [especially high systolic BP (SBP)] and HD or CVD (12–23). Although most of these studies indicated that hypertension is a risk factor for HD and CVD, some questioned the relationship between them (13, 18–20, 22, 23).

Risk factors for CBD (stroke) are classified as non-modifiable and modifiable factors. The American Stroke Association (ASA) (24) stated that nonmodifiable risk factors include age, family history, race, gender, prior stroke history, transient ischemic attack (TIA), and heart attack. Modifiable risk factors include high BP, smoking, diabetes, diet, physical inactivity, obesity, high blood cholesterol, artery disease, peripheral artery disease, atrial fibrillation, and sickle cell disease. Various studies have been conducted on CBD (25–31). In case of CBD, long-term rehabilitation frequently becomes a serious problem in addition to prevention, detections, and treatment (25, 26, 30).

Furthermore, CDC (32) described diabetes and high BP as two important risk factors for chronic KD (CKD). HD, obesity, family history of CKD, inherited kidney disorders, past kidney damage, and older age were other risk factors. The American Kidney Fund (33) mentioned that diabetes is a leading risk factor, and high BP and race or ethnicity are other important risk factors. The Kidney Foundation of Canada (34) also mentioned that diabetes, high BP, and a family history of KD are risk factors. Since KD, especially CKD, is a serious global burden (35), many studies on its risk factors (36–40) and treatment procedures have been conducted (41–46). CKD is a costly disease to treat. Liyanage et al. (42) reported that “renal replacement therapy (RRT), through either dialysis or renal transplantation, is a lifesaving yet high-cost treatment for people with end-stage kidney disease.” Although the results of cost-effectiveness analyses (47–51) suggest that renal transplantation is more efficient than dialysis, few renal transplantations have been performed in Japan. The number of kidney transplantations in the United States (52) was 23,642 in 2020. In contrast, the numbers in Japan (53) were 216 and 127 in the fiscal years 2019 and 2020, respectively, and the number of dialysis patients (54) was 347,641 at the end of 2020. Nawata and Kimura (55) reported that the medical cost of an individual with KD was 14.5 times higher than that of an individual without KD.

In the present study, the risk factors for HD, CBD, and KD were analyzed using the JMDC Claims Database (56), which includes 13,157,681 medical checkups performed on 3,233,271 individuals in Japan. The side effects of medications to control hypertension (antihypertensive medications), hyperglycemia (antihyperglycemic medications), and hypercholesterolemia (cholesterol medications) on these diseases were also evaluated.

The estimation results of Model A (HD model) are presented in Table 3. The gross percentage of HD patients in the following year was 0.55% among individuals without HD history. The estimate of Age was positive, and its t-value was quite large and significant at any reasonable significance level. The estimates for Female, Family, and t1 were negative and significant at the 1% level. The estimates for HbA1c, AST, U_Protein, and Weight_1 were positive and significant at the 1% level. The estimates of HDL, LDL, ALT, and Sleep were negative and significant at the 1% level. The estimates of CBD_t and KD_t were positive and significant at any reasonable level. The estimates of all dummy variables that represent medication use were positive, and their t-values were quite large and significant at any reasonable significance level. In particular, when antihypertensive medications were included, the estimated values were 0.657, 0.595, 0.793, and 0.784 for M_BP, M_BP&GL, M_BP&CH, and M_BP&GL&CH, respectively. These values were much larger than those obtained without antihypertensive medications.

Table 4 shows the results of the estimation of Model B (CBD model). The gross percentage of those with CBD in the following year was 0.23%. The estimate of Age was positive, and its t-value was quite large and significant at any reasonable significance level. The estimates for DBP, GGP, U_Protein, and Weight_1 were positive and significant at the 1% level. In contrast, the estimates of Speed and Sleep were negative and significant at the 1% level and that of LDL was significant at the 5% level. The estimates of HD_t and KD_t were positive and significant at any reasonable level. The estimates of all dummy variables that represent medication use were positive, and their t-values were quite large and significant at any reasonable significance level. In particular, when antihypertensive medications were used, the estimated values were much higher than those without antihypertensive medications.

The estimated results for Model C (KD model) are listed in Table 5. The percentage of patients with KD in the following year was 0.12%. Again, the estimate of Age was positive and its t-value was quite large and significant at any reasonable significance level. The estimate of t1 was positive and significant at the 1% level, while that of Female was negative and significant at the 5% level. The estimates of U_Protein and Weight_1 were positive and significant at the 1% level, and those of U_Sugar and Late_Supper were significant at the 5% level. The estimates of BMI, HbA1c, ALT, Smoke, Drink_Freq, and Activity were negative and significant at the 1% level. The estimates of HD_t and CBD_t were positive and significant at any reasonable significance level. As in the previous two models, the estimates of all dummy variables that represent medication use were positive and significant at any reasonable significance level. Especially, for the variables with antihypertensive medications, the estimates of M_BP, M_BP&GL, M_BP&CH, and M_BP&GL&CH were 1.01, 1.17, 1.30, and 1.27, respectively. These estimates were almost twice as large as those obtained without antihypertensive medications.

Table 6 summarizes the significant covariates other than disease histories and variables that represent taking medications in Models A (HD model), B (CBD model), and C (KD model). As expected, Age is a very important variable affecting these diseases. As shown in Figure 2, the odds ratios (ORs) comparing individuals aged 50–70 years were 2.09 with a 95% confidence interval (CI) of 2.01–2.18 for HD, 2.27 with a 95% CI of 2.12–2.41 for CBD, and 1.34 with a 95% CI of 1.22–1.45 for KD. Although Nawata (23) evaluated the risk factors for HD, the ORs and CIs were not evaluated. Since the probabilities of having these diseases in the next year are small, OR and CI are approximately equal to the probability ratio (PR) and its CI (57). The risk of HD and CBD in individuals aged 70 years is more than twice as large as those of individuals aged 50 years. The increase in risk was relatively small for KD. Female is significant at the 1% level in Model A but not significant at the 5% level in other models.

Weight_1 (recent weight change) and U_Protein (urine protein level) were significant at the 1% level in all models and were considered important variables. Figure 3 shows the ORs and 95% CIs of these variables for the models. For the dummy variable, the OR was calculated by comparing 0 and 1. For U_Protein, the majority of the values were 1 or 2, so the OR was calculated by comparing U_Protein = 1 and 2. For HD, the ORs of U_Protein and Weight_1 (95% CI) were 1.10 (1.0–1.13) and 1.20 (1.15–1.24.) For CBD, the ORs were1.13 (1.08–1.17) and 1.24 (1.17–1.31). For KD, the ORs were 2.23 (2.20–2.31) and 1.21 (1.11–1.30). Weight_1 increased the risk by approximately 20% for all diseases. U_Protein increased the risk by about 10% for HD and CD; however, it doubled the risk for KD. U_Protein is a risk factor with obvious mechanisms assumed (58) and people with high urine protein levels should recognize this fact, and special attention is necessary.

Another interesting finding is that among BP levels (SBP and DBP), only DBP in the CBD model was significant, and the estimates were not significant in the other models. This may raise questions regarding the 2017ACC/AHA guideline, which lowered the hypertensive criterion.

Figure 4 shows the ORs and 95% CIs for the variables of a history of other diseases. In the HD model, the ORs were 1.72 (1.54–1.89) and 1.71 (1.35–2.06) for CBD_t and KD_t. In the CBD model, the ORs were 2.15 (1.93–2.36) and 2.17 (1.61–2.74) for HD_t and KD_t. In the KD model, the ORs were 1.57 (1.34–1.81) and 1.58 (1.27–1.89) for HD_t and CBD_t. The risk of having CBD for individuals with HD and KD histories was especially high, more than double of those without disease histories, and special care is necessary for them (57).

All medications (antihypertensive, antihyperglycemic, and cholesterol-lowering) increased the risk of all three diseases. Figures 5–7 show ORs and 95% CIs, respectively. People often take two or three types of medications, and seven different variables are separately shown in the figures. These variables were divided into two groups. One group was treated with antihypertensive medications (i.e., M_BP, M_BP&GL, M_BP&CH, and M_BP&GL&CH), and the other was without antihypertensive medications (i.e., M_GL, M_CH, and M_GL&CH). In the HD model, the ORs (95% CI) were 1.93 (1.83–2.02), 1.81 (1.59–2.04), 2.21 (2.06–2.36), and 2.19 (1.93–2.45) for M_BP, M_BP&GL, M_BP&CH, and M_BP&GL&CH, respectively. In contrast, the ORs were 1.26 (1.10–1.43), 1.41 (1.31–1.52), and 1.29 (1.07–1.51) for M_GL, M_CH, and M_GL&CH. In the former group, the risks associated with antihypertensive medications were almost double compared with those without medications. In the latter group, the increments were much smaller, approximately 25%–40%.

In the CBD model, the ORs (95% CI) were 1.89 (1.75–2.04), 2.15 (1.77–2.53), 2.20 (1.98–2.43), and 2.17 (1.79–2.55) for M_BP, M_BP&GL, M_BP&CH, and M_BP&GL&CH, respectively. In contrast, the ORs were 1.47 (1.18–1.86), 1.66 (1.49–1.84), and 1.62 (1.22–2.02) for M_GL, M_CH, and M_GL&CH. In the former group, the risks associated with medications were almost double compared to those without medication, as in the case of heart disease. In the latter group, the increments were about 50%–65%. The results of this study suggest that antihypertensive medications increase the risks of HD and CBD, and the statement of Muntner et al. (11) is not supported.

Concerning the KD model, the ORs (95% CI) were 2.75 (2.46–3.93), 3.22 (2.44–4.01), 3.69 (3.20–4.18), and 3.59 (2.77–4.41) for M_BP, M_BP&GL, M_BP&CH, and M_BP&GL&CH, respectively. In contrast, the ORs were 1.76 (1.22–2.31), 1.61 (1.35–1.88), and 1.93 (1.17–2.68) for M_GL, M_CH, and M_GL&CH. In the KD mode1, although the CIs were larger, the increments in the risks seem much higher than those of the previous two diseases. The risks triple (or increase more) in the first group. The increments were about 60%–70% in the second group. Especially in the first group, even the lower bounds of the 95% CIs, the risks become 2.4–3.2 times larger than those without medications. This may imply that the negative effects of antihypertensive medications are much more serious in patients with KD.

The interactions between medications are important issues when individuals use two or more different types of medications. In the models without antihypertensive medications, the 95% CIs of M_GL and M_CH overlapped with the CI of M_GL&CH, and no significant interaction effect was observed. However, when antihypertensive and cholesterol medications were used simultaneously (M_ BP & CH), the risks were significantly higher (i.e., the 95% CIs did not overlap) than those using only antihypertensive medications (M_BP) in the HD and KD models.

Not only are the values of estimates of antihypertensive medications much higher than those of the other medications but the use of antihypertensive medications is also higher than that of other medications (antihypertensive medications: 11.18%, antihyperglycemic medications: 3.21%, and cholesterol medications: 7.46%). Here, we mainly discuss antihypertensive medications. For possible interpretations of the positive estimates, see Nawata (23).

Although there are some alternative therapies such as renal denervation (59), prescribing antihypertensive medications is a major therapy for hypertension. Over 200 drugs are available globally (60). Jackson and Bellamy (61) classified these medications into two group. One group is being those which directly or indirectly block the renin–angiotensin system (RAS) and the other group works through non-RAS pathways. The first group includes angiotensin-converting enzyme inhibitors (ACEIs), angiotensin-converting enzyme inhibitors, angiotensin receptor antagonists and direct renin inhibitors. The second group consists of adrenoceptor antagonists (β-blockers, α-blockers), calcium channel blockers, diuretics (thiazides, loop, potassium sparing/aldosterone antagonist), vasodilators, centrally acting agents, and ganglion block.

Laurent (62) also classified the medications into β-blockers, diuretics, ACEIs, angiotensin II receptor blockers (ARBs), calcium-channel blockers, and other classes. Although the information of individual prescription records is not available, ACIs and ARAs are most widely used antihypertensive medications in Japan (63). RAS is a regulator of blood volume, electrolyte balance, and systemic vascular resistance. For summaries of RAS in the management of hypertension, see Fountain et al. (64) and Lange-Jacobs et al. (65). RAS is an extreme complex system in playing hypertension, and the understanding of RAS has expanded tremendously over the last few decades (65). Crifciler and Haznedarouglu (66) outlined intersections of circulating and local angiotensin systems in the vascular pathobiological microenvironment of central nervous system.

Since antihypertensive medications are widely used, their mechanisms of action, effectiveness, proper usage methods, and side effects have been widely studied (67–76). Lithell (67) reported that “both β-blockers and thiazide diuretics worsened insulin resistance and deteriorated lipoprotein metabolism. …These data may explain the unexpectedly high incidence of diabetes among hypertensive patients and the poor effect on risk for coronary heart disease in intervention trials.” Santos et al. (68) did systematic reviews to identify the possible pharmacogenetic implications for RAS-blocker medications in the hypertension-CKD scenario. They described that studies focusing on CKD were scarce. The CDC (77) and International Diabetes Foundation (78) described KD as a serious diabetes complication.

Marcum and Fried (79) mentioned that antihypertensive medications are frequently used in older adults with CKD, and the most common adverse drug events (ADEs) with antihypertensive use include acute kidney injury and antihypertensives, which may lead to medication errors and ADEs. Even in the Systolic Blood Pressure Intervention Trial (SPRINT) (16), which was heavily weighted in the 2017ACC/AHA guideline, it was admitted that “Rates of serious adverse events of hypotension, syncope, electrolyte abnormalities, and acute kidney injury or failure, …, were higher in the intensive-treatment group than in the standard-treatment group.” They also reported the percentages of the serious adverse events in both groups. These findings are consistent with the results of the present study. Although Ptinopoulou et al. (80) wrote that “it is now common knowledge that adequate blood pressure control is the most important factor for the preservation of renal function, so every drug that effectively lowers hypertension is believed to be renoprotective,” the findings of the paper do not support this statement. The risks of hypertensive medications were three times or more higher than those without medications for KD. Any medication may have negative side effects. In particular, the risk of KD is much higher for individuals taking antihypertensive medications. These medications must be prescribed with caution to minimize any negative side effect, and further studies are needed on the risks of antihypertensive medications.

We obtained slightly different results from those of previous studies. One possible reason might be a sample selection bias. In this study, we analyzed the general population. In other studies, special individuals with a high risk of diseases were analyzed. The rate of having heart diseases in the next year is about 0.5%. If the sample size is 10 thousand, the expected number of new patients is just 50 per year and it is sometimes difficult to get statistically significant results. In the SPRINT (16), they selected 9,361patients (4,678: intensive treatment; 4,683: standard treatment) at high risk for cardiovascular events. They admitted that “The lack of generalizability to populations … is a limitation.” Over 2.8 million observations were used in the study and the generalizability problem might be improved. For the theoretical details, see Nawata and Kimura (19).

In this study, the risk factors for heart, cerebrovascular, and kidney diseases (HD, CBD, and KD) were analyzed using data from 2,837,334, 2,864,874, and 2,870,262 medical checkups obtained from the JMDC claims database. The data of individuals who had no history of each disease at year t and had information in the following year were analyzed using the logit models. Among the health-related information, age was a very important risk factor. The histories of other types of diseases were also very important factors, and the risks of HD, CBD, and KD would increase when individuals had histories of other diseases.

Among other factors, urine protein levels and recent large weight changes were very important for all three diseases. For KD, the risk would be more than double for individuals with high urine protein levels, and these individuals should be aware of this fact to prevent KD.

The side effects of the medications, including interactions, were also evaluated. Taking antihypertensive medications increased the risk of disease. The negative side effects were especially severe when individuals were taking antihypertensive drugs. Antihypertensive medications are widely used. Although the 2017ACC/AHA guideline lowered the hypertensive criterion, special care and additional studies are necessary to prescribe these medications, particularly antihypertensive medications so as to minimize the negative side effects of medications.

As the dataset comprises the health checkups of workers in Japan, it does not include individuals aged 76 and above. Age is a very important factor in these diseases, and the risk of these diseases increases with age. Although no precise data are not available, it is indicated that adequate care and consideration should be given to antihypertensive medications for the elderly. It is also necessary to collect data on older adults. The dataset only contained information obtained in Japan. The Japanese are ethnically homogeneous, and potential ethnic effects on the diseases are not evaluated. This might be an advantage to examine factors such as effects of medications because underlying genetic influences might be small. However, we may get different results in other countries, and that is a limitation of this study. Further studies are needed for generalization of this study. These will be the subjects of future studies.