Association of Antihypertensive Drug-Related Gene Polymorphisms with Stroke in the Chinese Hypertensive Population

Background Antihypertensive therapy is crucial for preventing stroke in hypertensive patients. However, the efficacy of antihypertensive therapy varies across individuals, partially due to therapy-related genetic variations among individuals. We investigated the association of antihypertensive drug-related gene polymorphism with stroke in patients with hypertension. Methods Demographic information, medication, and outcome data were obtained from a hypertensive patient management system, and a PCR fluorescence probe technique was used to detect 7 gene polymorphic loci (CYP2D6∗10, ADRB1, CYP2C9∗3, AGTR1, ACE, CYP3A5∗3, and NPPA), and these loci were compared between patients with and without stroke. Logistic regression was performed to analyze the association of these genetic variations with stroke risk in hypertensive patients while controlling for potential factors. Results The prevalence of stroke in the hypertensive population in Changsha County of Hunan Province was 2.75%. The mutation frequencies of ADRB1 (1165G > C), CYP2D6∗10, CYP2C9∗3, AGTR1 (1166A > C), ACE (I/D), NPPA (2238T > C), and CYP3A5∗3 were 74.43%, 57.23%, 4.26%, 5.71%, 31.62%, 1.17%, and 69.58%, respectively. Univariate analysis revealed that ADRB1 polymorphism was associated with stroke (χ2 = 8.659, P < 0.05), with a higher stroke risk in the CC group than in the GC and GG groups (GC + GG). Multivariate unconditional logistic regression analysis showed that the CC genotype in ADRB1 (vs. the GC + GG genotype) was associated with an increased risk of stroke [odds ratio (OR) = 1.184, P<0.05] in hypertensive patients. No association was observed between CYP2D6∗10, CYP2C9∗3, AGTR1 (1166A > C), ACE (I/D), CYP3A5∗3, and NPPA (2238T > C) polymorphisms and stroke. Conclusions ADRB1 (1165G > C) gene polymorphism is associated with the risk of stroke in Chinese hypertensive patients. The CC genotype is correlated with a higher risk of stroke than the GC + GG genotype.


Introduction
Stroke, which is the second leading cause of death globally, is a chronic noninfectious disease hampering public health [1].Compared to that in 1990, the absolute number of stroke events worldwide increased by 70.0% in 2019, and the number of deaths increased by 40%.Currently, stroke is the leading cause of death and disability among adults in China; its morbidity, mortality, and number of cases are the highest worldwide [2,3].
Hypertension is the most important and modifable high-risk factor for stroke development [4][5][6].Te World Health Organization (WHO) estimates that approximately 54% of all stroke cases worldwide can be attributed to hypertension [7].Hypertension control could signifcantly reduce the incidence of stroke [8].Te drugs for the treatment of hypertension are mainly divided into fve categories: beta receptor blockers (BBs), angiotensin-II receptor antagonists (ARBs), angiotensin-converting enzyme inhibitors (ACEIs), calcium channel blockers (CCBs), and diuretics.However, community attainment of hypertension treatment remains dismal, with less than a quarter being controlled [9,10].Uncontrolled hypertension may signifcantly contribute to high mortality from stroke [11].
Diferences in these gene polymorphisms in response to antihypertensive therapy have been described previously in the literature.Certain genotypes exhibit a greater reduction in blood pressure (BP) and/or long-term beneft with antihypertensive drug use [15].In particular, the CC genotype of ADRB1 (1165G > C), DD genotype of ACE (I/D), C allele carrier of AGTR1 (1166A > C), and NPPA (2238T > C) are reported to be associated with a signifcant reduction in BP [16][17][18][19].Mutations in cytochrome P450 enzyme family members, such as CYP2D6, CYP2C9, and CYP3A5, mainly afect the activities of metabolic enzymes.Patients using antihypertensive drugs and having an increased active CYP phenotype may have increased metabolic capacity; in contrast, those with an inactive CYP phenotype may display increased clinical efects and bradycardia following the use of metoprolol and other antihypertensive drugs [20].A previous study reported that the CYP2C9 * 3 allele may be associated with a decrease in the activity of the CYP2C9 enzyme in the Asian population [21], while CYP3A5 * 3 is correlated with efective control of BP with amlodipine (CCB) in hypertensive patients following renal transplantation via decreasing the activity of the CYP3A5 enzyme [22].Te association of the CYP2D6 mutation with the outcome of antihypertensive therapy with BBs is controversial.CYP2D6 * 10 signifcantly afected the pharmacokinetics of metoprolol in the Chinese population [23]; however, no signifcant efect on the outcome of carvedilol and propranolol therapy was observed [24,25].
are associated not only with BP reduction but also with disease prognosis.A few previous studies revealed that a mutation in ADRB1 (1165G > C) was associated with the survival of patients with heart failure by increasing the dose of BBs [26].In patients with chronic heart failure and the ADRB1 CC genotype, the use of BBs also appears to reduce the risk for cardiovascular events [26,27].However, to the best of our knowledge, no studies have focused on the relationship between all above-mentioned genes and stroke.A community registry of hypertensive patients was started in Changsha County, Hunan Province, in 2013; this system registers the BP treatment and outcomes of hypertensive patients through follow-ups.In 2017, a test of the seven genes polymorphisms mentioned above was conducted in hypertensive patients.Our study aimed to explore the association between diferent antihypertensive-related genotypes and stroke.It would provide some evidence for preferential selection of antihypertensive medication in the hypertensive population in Changsha County, Hunan Province.

Study Population.
Te study population was selected from hypertensive patients registered in the hypertension management system at each township health center and street health service center in Changsha County, Hunan Province, between January 1, 2013, and January 1, 2018.
Inclusion criteria for study participants: all the hypertensive patients registered in the management system; these patients were enrolled based on the criteria of the Chinese Guidelines for the Prevention and Treatment of Hypertension (2010 Revision) [28].(1) participants had a systolic blood pressure (SBP) ≥ 140 mmHg (1 mmHg � 0.133 kPa) or a diastolic blood pressure (DBP) ≥ 90 mmHg three times on diferent days; (2) participants were previously diagnosed with hypertension; and (3) participants were currently taking BP medications.Exclusion criteria for study participants: (1) participants registered in the system but with no record of BP; (2) participants without any follow-up records; and (3) participants under the age of 18 years.

Data Source.
All patient data were obtained from the same above-mentioned hypertension management system.Tis system is a part of the National Basic Public Health Service Project Management Information System, which records the medical information of all registered and followup hypertensive patients.According to the basic public health service guidelines, hypertensive patients were followed up face-to-face four times a year, and they had to undergo physical examinations yearly.Family doctors and public health personnel in the community health service centers were responsible for registering, following up, and managing patients with hypertension within their jurisdiction.Te Centers for Disease Control and Prevention were responsible for developing technical programs, personnel training, and quality control of follow-up data.All medical follow-up information of the patients must be faithfully recorded in the system database by the physicians monitoring the patients.Strokes and other illnesses in follow-up patients must be documented in detail; stroke and other cardiovascular events must be diagnosed and confrmed by a high-level hospital.
Tis study was approved by the Ethics Committee of Xiangya School of Public Health, Central South University (approval number: XYGW-2022-73), and informed consent was taken from all individual participants.

Data Collection.
Demographic information, medication, and outcome data were derived from the basic patient information module, physical examination information module, drug exposure module, and follow-up module in the hypertension patient management system in Changsha County, Hunan Province.extending 40 s at 72 °C; a total of 40 cycles; and the fnal extension step at 72 °C for 5 min.GenePix4 100A scanner was used to conduct quantitative analysis of the scan results with its own system, GenePix6.0,and the analysis software was used to interpret the genotype according to the set Cuf value.For genes and related polymorphisms, rs number, and interested genotypes, refer to the supplementary documents (Table S1).

Data Cleaning and Merging
2.4.1.Data Cleaning.Tis study included data derived from the basic information module with one record for each participant and from the physical examination follow-up and medication modules, generating multiple records for each participant.Terefore, we carried out data cleaning according to the following principles: (1) For the physical examination module, height, weight, BP, and positive physical examination results (such as stroke, coronary heart disease, and kidney disease) were mainly extracted.
(2) For follow-up and medication modules, the inpatient and outpatient diagnoses, the patient's main antihypertensive medication, and the duration of medication use were extracted.Te antihypertensive drugs were classifed into fve categories: BBs, ARBs, ACEIs, CCBs, and diuretics.CCBs include dihydropyridine and nondihydropyridine drugs.Accumulated days of each antihypertensive drug taken by the patients were calculated according to the medication status with and without specifc medicine, and compliance was recorded at each follow-up visit.If patients complied well with antihypertensive drugs, the duration of drug exposure was considered as 120 days during one follow-up period based on 4-month visit intervals.In the case of poor compliance of patients with drugs, the number of days of drug use during the follow-up period was registered or calculated by the time between the start and discontinuation of drug use.(3) Poststroke medication exposure was not analyzed.

Data Merge.
Basic information, BP at each examination and visit, antihypertensive drug use, stroke, and other conditions were combined with "resident ID card" as the unique identifcation number and name for the checking information.

Stroke Defnition.
In this study, stroke was defned as follow-up and physical examination records containing "stroke," "cerebral hemorrhage," "cerebral infarction," or records with diagnosis codes (ICD-10) containing I60-I64.Te cases extracted from the follow-up records require the appropriate diagnostic basis (inpatient treatment in hospitals at or above the county level or with a high-resolution imaging diagnosis).If a "stroke" record with unique identity was detected in the follow-up or physical examination record, it will be recorded as one patient.If "stroke" records with the same identity were detected both in the follow-up and physical examination records, they will be recorded as one patient only.

Body Mass Index Classifcation.
Body mass index (BMI) was calculated based on the height and weight of patients at their earliest registration dates and was divided into three categories: <18.5, 18.5∼24, and >24 kg/m 2 .

Age Classifcation.
According to the age at the earliest registration dates in the system, participants were divided into three age groups: 18-45, 45-60, and >60 years.

Relevant Family History and Disease Incidence.
Family history was defned as the patient's immediate family members with hypertension, stroke, diabetes, or other diseases.Presence of coronary heart disease, COPD, diabetes, or tumor refers to the occurrence of these diseases in International Journal of Hypertension the patients; the related data were generally obtained from the registration inquiry information belonging to the basic information data.

Control of SBP and DBP.
According to the hypertension prevention and treatment guidelines, control of SBP and DBP is defned as follows: (1) general hypertension patients with BP of <140/90 mmHg, (2) hypertension comorbid with diabetes or heart failure with BP of <130/80 mmHg, (3) hypertension comorbid with coronary heart disease or chronic kidney disease with BP of <140/90 mmHg, and (4) older patients aged >65 years with BP of <150/90 mmHg.If the above criteria were met, BP was considered to be under control; otherwise, it was not controlled.
2.5.6.History of Drug Exposure.According to the WHO [29] and Chinese guidelines [30], if the antihypertensive efect is not good or does not achieve the goal level after 4 weeks of initial treatment, a combination of medications or a change of medication is required.Terefore, we defned a history of use of an antihypertensive drug as the drug being used for more than 28 days.
Administering antihypertensive drugs (BBs, ARBs, ACEIs, CCBs, and diuretics) refers to a patient's taking a specifc antihypertensive drug for several days.According to Guidelines for Rational Drug Use of Hypertension in China (Second Edition) [30], the use of BBs, ARBs, ACEIs, CCBs, and diuretics should be changed or used in combination therapies after four weeks of administration.Terefore, we defned the exposure time of BBs, ARBs, ACEIs, CCBs, or diuretics as more than 4 weeks.
2.6.Statistical Analysis.SPSS 22.0 software and R Studio were used for data analyses.Quantitative data and categorical data were expressed as mean and standard deviation (SD) and as frequency or percentage, respectively.A chisquare test was performed to compare the genotypes and allele frequency and detect Hardy-Weinberg equilibrium (HWE).Logistic regression was used to determine the association of the polymorphisms of each antihypertensive drug-related gene with stroke, considering the possible confounding factors, including general demographic information, the presence of other main diseases, family history of disease, and antihypertensive drug use.All hypothesis tests were two-tailed, and P < 0.05 was considered statistically signifcant.A further subgroup analysis was performed, considering the clinical heterogeneity of hypertension patients.

General Information.
A total of 33,210 participants were registered in the system.We excluded 90 patients who were enrolled in the system but lacked genetic information and 3,458 patients who lacked BP records and follow-up records.One patient under 18 years was excluded.Finally, 29,661 patients with hypertension were included in the follow-up analysis (Figure 1); among them, 816 (2.75%) had stroke.Signifcant diferences were observed between the stroke patients and the nonstroke patients in terms of sex, age, BMI, educational level, presence of coronary heart disease, presence of COPD, presence of tumor, family history of stroke, family history of hypertension, and the use of ACEIs, diuretics, or CCBs (P < 0.05).In contrast, no signifcant diferences were observed between these two groups in terms of resident type, marital status, control of SBP, control of DBP, presence of diabetes, family history of diabetes, ARBs exposure, and BBs exposure (P > 0.05) (Table 1).2).Hardy-Weinberg balance test was conducted on the distribution of antihypertensive drug-related genotypes in the control group, and no signifcant diference was observed in ADRB1 (1165G > C), CYP2C9 * 3, ACE (I/D), and CYP3A5 * 3 polymorphism loci (P > 0.05).Te polymorphism loci of CYP2D6 * 10, AGTR1 (1166A > C), and NPPA (2238T > C) did not conform to gene balance (P < 0.05) (Table 3).Univariate analysis revealed that the CC genotype of ADRB1 (1165G > C) was strongly associated with a higher stroke risk than the GC and CC genotypes (χ 2 = 8.659, P < 0.05) (Table 3).Tis association remained signifcant after multivariate unconditional logistic regression analysis was performed adjusting for age, sex, BMI, type of usual residence, education, marital status, control of SBP, control of DBP, presence of disease, family history, and antihypertensive drug history.Tese observations revealed that the risk of stroke was 1.184 times higher in CC carriers with the dominant gene than in GC + GG genotype patients (95% confdence interval (CI): 1.026-1.365,P < 0.05) (Table 4).No association was observed between CYP2D6 * 10,CYP2C9 * 3, AGTR1 (1166A > C), ACE (I/D), CYP3A5 * 3, and NPPA (2238T > C) polymorphisms and stroke (Tables 3 and 4).

Subgroup Analysis of the Relationship between Antihypertensive Drug-Related Genes and Stroke in Hypertensive
Patients.To further understand the specifc association of ADRB1 (CC vs. GC + GG) with stroke in diferent subpopulations, we conducted regression analysis by dividing the study subjects into diferent subgroups according to the presence of diabetes mellitus, presence of coronary heart disease, presence of COPD, presence of tumor, and whether they were taking BBs, ARBs, ACEIs, CCBs, and diuretics.Te results revealed that CC (vs GC + GG) of the ADRB1 (1165G > C) gene was associated with stroke occurrence in nondiabetic and noncoronary populations of hypertensive patients (OR: 1.20, 95% CI: International Journal of Hypertension

Discussion
Tis study revealed that the prevalence of stroke in the hypertensive population in Changsha County of Hunan Province was 2.75%, which was signifcantly lower than the 10.04% reported in hospitalized hypertensive patients in the Anhui Province [31].Te diference may be a result of differences in the study population, with the hospitalized population having more difculty in controlling BP and higher-grade hypertension than the community population.
Furthermore, the BMI and age were lower in the present study than in the previous one.Te prevalence of stroke in male patients was higher than that in female patients possibly because female patients had better hypertension control and a lower incidence of stroke than male patients [9].
Our study showed that the mutation frequencies of antihypertensive drug-related genes varied from 5.71% to 74.43%.Te mutation rate of the ADRB1 (1165G > C) gene (74.43%) was slightly higher than that in Chinese patients with essential hypertension reported previously (59.8%) [32], while the mutation rate of CYP3A5 * 3 (69.58%)was slightly lower than that reported previously (79.5%) [33].Te diference may be explained by the fact that the present study recruited hypertensive patients in the community, while the previous research subjects were hospitalized hypertensive patients [32]   1 † * P < 0.05.† Reference.‡ Adjusted for age, sex, BMI, type of usual residence, education, marital status, control of SBP, control of DBP, presence of diabetes, presence of coronary heart disease, presence of COPD, presence of tumor, and family history related to hypertension, stroke, diabetes as covariates.§ Adjusted for age, sex, BMI, type of usual residence, education, marital status, control of SBP, control of DBP, presence of diabetes, presence of coronary heart disease, presence of COPD, presence of tumor, family history related to hypertension, stroke, diabetes, and whether taking BBs, ACEIs, ARBs, CCBs, and diuretics as covariates.
International Journal of Hypertension or healthy individuals [33].Te frequency of CYP2D6 * 10, CYP2C9 * 3, and AGTR1 (1166A > C) was generally consistent with that reported for Chinese hypertensive patients in previous studies [21,34].In addition, we observed that CYP2D6 * 10, AGTR1 (1166A > C), and NPPA (2238T > C) genes do not conform to the Hardy-Weinberg equilibrium test.Another study conducted in Chinese Han hypertensive patients also found that CYP2D6 * 10 was not in HWE [16]; this might be because CYP2D6 * 10, AGTR1 (1166A > C), and NPPA (2238T > C) genes were associated with the development of hypertension [23,35,36].Studies revealed that alleles associated with disease etiology often deviate from the expected allele or genotype frequency [37].
Our results indicated that the CC genotype of ADRB1 (1165G > C) is a risk factor for stroke in hypertensive patients, and the risk of stroke was 1.184 times higher for the CC genotype than for the GC + GG genotype (95% CI: 1.026-1.365).Studies had established that ADRB1 (1165G > C) dominant allele C was a risk factor for hypertension [38][39][40][41].It is reasonable to assume that CC may also contribute to the risk of developing stroke in hypertensive patients.In addition, previous studies indicated that mutation of ADRB1 (1165G > C) was associated with shorter sleep duration in humans and mice; carriers of the mutation sleep two hours less per day on average than nonmutation carriers [42], and shorter sleep duration was an independent risk factor for future stroke events in hypertensive patients [43].Terefore, ADRB1 (1165G > C) might also afect the occurrence of stroke by afecting sleep.Our study was inconsistent with the results of a study on ADRB1 (1165G > C) polymorphism and ischemic stroke in North India, which recruited 224 patients and 224 age-and sexmatched controls and discovered no association between ADRB1 (GC + GG vs. CC) and overall ischemic stroke [44].Te following reasons can explain the diference: (1) Te diference of interested diseases: Our study recruited all patients diagnosed with ICD codes ranging from I60 to I64, while the Indian study only included patients with ischemic diseases (I63); (2) heterogeneity of design methods: Tis study is a cross-sectional study and the previous one is case-control study; and (3) the sample diference: Tis study had a larger number of subjects derived from the community, and the case control had a smaller sample size and was hospital-based.Although both studies had some bias, we believe that genes are generally associated with stroke.
To further analyze the relationship between ADRB1 (1165G > C) polymorphism and stroke, we performed a subgroup analysis.Te results indicated that ADRB1 (1165G > C) polymorphism was signifcantly associated with stroke risk in patients with noncoronary heart disease and nondiabetic diseases, while not signifcantly associated with stroke risk in patients with coronary heart disease and diabetes.Tis fnding suggests that hypertensive patients  [31,45,46], might afect stroke development and conceal the association of ADRB1 (1165G > C) gene polymorphism with stroke.Similarly, a meta-analysis revealed no signifcant association of ADRB1 (1165G > C) polymorphisms with cardiovascular events in patients with coronary artery disease [47].
Te result of subgroup analysis revealed that polymorphisms in the ADRB1 (1165G > C) gene were consistently associated with stroke in hypertensive patients regardless of BBs use.Te OR values were 2.426 (95% CI: 1.077-5.466)and 1.167 (95% CI: 1.010-1.348) in hypertensive CC and GC + GG carriers, respectively.Numerous studies have demonstrated that the CC genotype is associated with a better antihypertensive response to metoprolol treatment in diferent racial populations of healthy volunteers and hypertensive patients [26,48].A recent study assessed the role of ADRB1 (1165G > C) gene polymorphisms on BBs response in a population with heart failure, revealing that the patients with the CC genotype seemed to receive the most beneft from a high BBs dose.However, our study showed the opposite result.Te discrepancy could be explained by the following reasons: (1) BBs are the frst-line pharmacological treatment for hypertensive patient comorbid with other conditions, such as heart failure and coronary disease, but not the frst-line treatment for simple blood pressure-elevated patients.Terefore, for those simple blood pressure elevated patients with the CC genotype, BBs treatment was prescribed until "trial and error" treatment failed or developed with other conditions, the beneft from BBs prescription of CC genotype patients may have been delayed, and lead to a higher risk of stroke.(2) Although we excluded the data of poststroke drug usage, the patients may have experienced a prolonged period of uncontrolled BP before the stroke, and during this period, BBs may have been used.A study showed that there may be a higher risk of stroke due to greater variability in BP when treated with BBs [49], increasing the risk among patients with BBs.Our results indicated that ADRB1 (1165G > C) was not associated with stroke occurrence in hypertensive patients without diuretics, and ADRB1 (1165G > C) was associated with stroke risk in hypertensive patients with diuretics, possibly because the patients without diuretics were those with a shorter course of the disease and a lower BP level.Te antihypertensive efect of diuretics was relatively weak when used as antihypertensive drugs in hypertensive patients for a long time [50].Terefore, long-term use of diuretics may lead to poor BP control and increase the risk of stroke.
We did not establish an association of CYP2D6 * 10, CYP2C9 * 3, AGTR1 (1166A > C), ACE (I/D), CYP3A5 * 3, and NPPA (2238T > C) gene polymorphisms with stroke susceptibility.Te result of a study conducted in India was consistent with our observations, in that no signifcant association of ACE (I/D) gene polymorphisms with ischemic stroke exists in hypertensive patients [51].Te secondary C allele of the NPPA (2238T > C) gene is associated with a higher risk of stroke [52]; however, no association between the NPPA (2238T > C) gene and stroke susceptibility was detected in other studies [19].Regarding the AGTR1 (1166A > C) gene, studies revealed that its C allele was associated with increased BP in hypertensive patients [35]; currently, there are no studies on its relationship with stroke in hypertensive patients.Terefore, the association of CYP2D6 * 10, CYP2C9 * 3, AGTR1 (1166A > C), ACE (I/D), CYP3A5 * 3, and NPPA (2238T > C) gene polymorphisms with stroke in hypertensive patients needs further studies.
Tis study has some limitations.First, stroke is a disease with complex pathogenesis; hemorrhagic stroke and ischemic stroke have diferent mechanisms, conditions, and prognoses.Te data used in this study could not distinguish between ischemic and hemorrhagic stroke.Terefore, the study could not analyze the diferences between these two types of stroke.Second, all the participants were Han people, and there may be genetic variations among different ethnic groups.In future studies, we will verify the results of this study with data from other groups.Tird, the stroke cases identifed in the hypertension management system were mainly prevalent cases; therefore, the study was inclined to generate a Neiman bias.As we explained in the previous paragraph, although we excluded the poststroke data, the patients may have experienced a long period of uncontrolled BP before the stroke, and during this period, medications and lifestyles may have changed, leading to a bias.Terefore, new cases should be selected, and more accurate exposure assessment methods are needed in future studies.However, our study systematically investigated the association of antihypertensive drugrelated gene polymorphisms and the occurrence of stroke in hypertensive patients by including a large number of participants to obtain reliable data through a hypertension management system and adjusting for possible nongenetic factors that may infuence the occurrence of stroke, such as sex, age, marital status, education level, BMI, family history, and antihypertensive medications.

Table 1 :
Demographic characteristics and associated risk factors of patients with stroke and healthy controls.

Table 2 :
Genotypes and allele frequency distribution in hypertensive population.

Table 3 :
Distribution of polymorphisms in genes related to antihypertensive drugs.

Table 4 :
Association between antihypertensive drug-related genes polymorphism and stroke occurrence in hypertensive patients (controls � 28846 and stroke � 816).

was not analyzed because the sample size was too small
In patients only with elevated BP, the CC genotype was the main risk factor for stroke; in hypertensive patients with other comorbid conditions, many nongenetic factors related to coronary heart disease or diabetes, such as smoking, physical activities, and dietary factors *It