The Journal of Applied Research • Vol. 4, No. 2, 2004 293

events independent of conventionalpredictors. The b3-adrenergic receptor(ADRB3) gene polymorphism is alsoassociated with elevated blood pressure(BP) and cardiovascular disorders. Inthis study, we examined the impact ofLVH (diagnosed by an electrocardio-graph) and ADRB3 polymorphism onresponse to anti-hypertensive therapy.

Methods: We treated a hypertensivecohort with benazepril for a period of15 days, and evaluated the drug effectby BP drop (∆BP). We performed stan-dard ECG measurement on each sub-

Electrocardiographic LeftVentricular Hypertrophy, b3Adrenergic ReceptorPolymorphism and Endpoint ofACEI Treatment of HypertensionKe Hao, ScD*†

Shaojie Peng, MD*

Aiqun Huang, MD‡

Xiumei Hong, MD*‡

Yan Zhang, MD, PhD§||

Xiaotao Zhao, MD||

Houxun Xing, MD*

Jianping Li, MD, PhD*

Alayne Ronnenberg, ScD†

Xin Xu, MD, PhD†

Tianhua Niu, ScD†

Xiping Xu, MD, PhD*†‡

*Institute of Biomedicine, Anhui Medical University, Hefei, China†Program for Population Genetics, Harvard School of Public Health, Boston, USA‡School of Life Science, University of Science and Technology in China, Hefei, China§Center for Eco-Genetics and Reproductive Health, Department of Genetics and Cell Biology,Peking University School of Medicine, China||Department of Cardiovascular Diseases, Peking University First Affiliated Hospital, Beijing,China

KEY WORDS: hypertension; electro-cardiography; genetic polymorphisms;ACE inhibitors

AbstractObjectives: Current treatment strategiesfor essential hypertension are effectiveonly in a portion of patients. Many fac-tors influence the therapeutic effect,including the status of left ventricularhypertrophy (LVH) and genetic varia-tions. LVH influences the intensity ofleft ventricular depolarization and is arisk factor for many cardiovascular

ject and determined ADRB3 genotypesexperimentally. Smoothing plots werefirst used to visualize the relationship,and additive models were fitted to inves-tigate the role of LVH and ADRB3polymorphism as well as their interac-tions with the treatment effect.

Results: Electrocardiographic LVH waspositively associated baseline BP, andnegatively associated with ∆BP. Therelationship was in a linear pattern. Theassociation is greatly modified byADRB3 genotype. In ADRB3 wild-type patients, the effect size increased,however, among ADRB3 mutant sub-jects, the association was attenuated andremained statistically non-significant.

Conclusion: Our results suggested thepredictive value of electrocardiographi-cally diagnosed LVH to the anti-hyper-tensive treatment effectiveness, andpointed out the importance of gene-environment interaction. Our findingcould have significant clinical implica-tions to optimize the therapeutic doserange based on the ECG indexes.

INTRODUCTIONPrimary hypertension affects more than20% of white adults, significantlyincreasing the risk of heart attack andstroke.1,2 Angiotensin convertingenzyme inhibitors (ACEI), such asbenazepril, are among the most widelyprescribed anti-hypertensive medica-tions.3,4 Despite their popularity, signifi-cant inter-individual variations in clinicalresponse to ACEI have been reported,5

which suggest that environmental orgenetic factors influence the drugs’effectiveness.

The Sokolow-Lyon Index (SLI) ofan electrocardiogram (ECG), defined asthe sum of RV5 and SV1, reflects theintensity of left ventricular depolariza-tion and hypertrophy.6-8 Subjects with ahigh SLI value are classified clinically as

having electrocardiographic left ventric-ular hypertrophy (LVH).6-8 LVH can beconsidered an adaptive response to theincreased afterload in hypertension,9,10

as well as a result of obesity or stimula-tion by angiotensin II or insulin.10

Several studies have suggested that LVHis a surrogate endpoint for many cardio-vascular events, independent of all con-ventional predictors.10-12 Despite therelation between LVH and hyperten-sion, the predictive value of LVH on theendpoint of anti-hypertensive therapyhas not been systematically addressed.

The b3-adrenergic receptor(ADRB3) is mainly expressed in adi-pose tissue and contributes to popula-tion variations in energy expenditureand body fat distribution.13 A missensemutation of the gene, resulting in substi-tution of tryptophan with arginine atcodon 64 (W64R), has been reported.14-16

This single nucleotide polymorphism(SNP) has been associated with insulinresistance, obesity, hyperinsulinemia, andhypertension,14-19 which could con-tribute to the development of LVH andcardiac dysfunction.20-22 It is possible,therefore, that the therapeutic effect ofACEI treatment varies according toADRB3 genotype and LVH status.

In this study, we treated a hyperten-sive cohort with benazepril for a periodof 15 days and evaluated the drug’seffectiveness in lowering blood pressure.The aim of the study was to investigatewhether ECG-diagnosed LVH is a sig-nificant predictor of blood pressuredecrease (∆BP) after drug therapy andwhether the ADRB3 W64R genotypemodifies the predictive effects of the ini-tial LVH on the decrease in BP.

METHODS AND MATERIALSStudy Site and PopulationThe study was conducted in HuoqiuCounty, Anhui Province, China fromJuly 2000 to May 2001. Huoqiu is locat-ed in the western part of Anhui

Vol. 4, No. 2, 2004 • The Journal of Applied Research294

province, among the plains and foothillsof Dabie Mountain. The residents speaka unique dialect, and the population hasexisted for hundreds of years with limit-ed migration and admixture. Subjectswho met the following criteria wereenrolled in the study: (1) diagnosed withdiastolic blood pressure (BP) of 90 to120 mmHg or systolic blood pressure of140 to 200 mmHg on three separateoccasions; (2) aged between 21 and 65years; and (3) lived in Houqiu Countyfor at least 2 years. To avoid misclassifi-cation and reduce the risk of severeadverse responses, we excluded patientswho (1) were diagnosed with secondaryhypertension; (2) were receiving antihy-pertensive treatment or had taken oth-ers drugs within 2 weeks prior to theexamination; (3) had a body mass index(BMI) > 33 kg/m2; or (4) had been diag-nosed with another severe disease ordisorder. Each eligible hypertensivepatient who was willing to participate

provided written consent. The institu-tional review board of Anhui MedicalUniversity and the Harvard UniversitySchool of Public Health approved thestudy procedure. The investigation con-forms with the principles outlined in theDeclaration of Helsinki.

Blood Pressure and ECG MeasurementsSupine blood pressure was measured at9:00 AM after 60 minutes of resting in asupine position. No alcohol, cigarettes,coffee or tea were taken for one hourprior to the measurements. Systolicblood pressure was defined as Korotkoffphase I (appearance of sound), and dias-tolic pressure as Korotkoff phase V (dis-appearance of sound), respectively.Three consecutive readings were takenon the same arm, with 30 secondsbetween each reading. The average ofthe 3 readings was used in our analysis.A standard 12-lead ECG examinationwas then performed on each participant.

The Journal of Applied Research • Vol. 4, No. 2, 2004 295

Table 1. Clinical Characteristics and ADRB3 Genotype of Study Cohort (n = 924)*

Mean ± SD

Age, y 48.4 ± 8.0

Height, cm 161.5 ± 8.0

Weight, kg 65.8 ± 12.0

Body Mass Index, kg/m2 25.2 ± 3.9

Blood pressure, mmHg

Systolic at baseline 149.4 ± 14.9

Diastolic at baseline 93.4 ± 8.5

Systolic after treatment 139.5 ± 17.3

Diastolic after treatment 88.2 ± 11.0

Sokolow-Lyon Index (mV ×¥ 10) 23.4 ± 7.7Percent

Ever anti-hypertensive drug use, % 14.0 %

Female, % 52.0 %

Former cigarette smoker, % 6.4 %

Current cigarette smoker, % 20.3 %

Former alcohol drinker, % 5.8 %

Current alcohol drinker, % 25.9 %

ADRB3 R64 Allele Frequency, % 16.3%

* ADRB3 indicates β3-adrenergic receptor; and SD, standard deviation.

Questionnaire After blood pressure was taken, patientswere interviewed using a standardizedquestionnaire. The following informa-tion was collected: (1) social-economiccharacteristics (age, gender, educationlevel, marital status, birthplace, familysize, household income, etc.); (2) person-al health history, including medicationuse, allergies, and history of hospitaliza-tions; (3) occupation history, includingcurrent and former jobs, work environ-ment, and possible occupational expo-sures; (4) lifestyle characteristics,including active/passive smoking, alcoholuse, and tea consumption; and (5)dietary factors, including intake of meat,fruit, vegetables, rice, and other foods.Body weight and height were also meas-ured to the nearest 0.1kg and 0.1cm,respectively, with light clothing and noshoes.

Benazepril TherapyAll qualified participants were asked totake one 10-mg tablet of benazepril(Ciba-Geigy, Beijing, China) everymorning for 15 consecutive days. Thefirst dose was taken at the study centeron day 1, and subsequent doses weretaken at the patient’s home for the next14 days. The subjects visited the studycenter every 3 days for follow-up, whichincluded measurement of blood pressureand heart rate and documentation ofother medications or therapies received

and any adverse reactions. On the 16thday, patients returned to the study cen-ter at 8 AM after fasting overnight. Theyagain rested in supine position for onefull hour, and supine blood pressure wasmeasured at 9 AM. Patients who report-ed taking other medicine during thestudy were excluded from the analysis.

Phlebotomy and DNA ExtractionBlood was drawn via venipuncture fromthe median cubital vein using a BDSafety-Lok blood collection set (BectonDickinson, Franklin Lakes, NJ) into a 10mL vacutainer (Becton Dickinson,Franklin Lakes, NJ) containing EDTA.The blood was kept on ice and subse-quently centrifuged for 10 minutes at4,000 × g at 4˚C. DNA was extracted atthe Institute of Biomedicine, AnhuiMedical University using the methodspreviously described. 23

GenotypingThe ADRB3 polymorphism was geno-typed using a PCR restriction fragmentlength polymorphism (RFLP) method.The genomic sequence flanking thepolymorphism was PCR-amplified usingprimers 5’-CAATACCGCCAA-CACCAGTGGG-3’ and 5’-GGTCATG-GTCTGGAGTCTCG-3’ in a volume of10 µL containing 30 ng of genomicDNA, 2.5 mM MgCl2, 200 µM of eachdeoxynucleotide triphosphate, 200 nMof each primer, 0.5 unit of Taq DNA

Vol. 4, No. 2, 2004 • The Journal of Applied Research296

Table 2. Association of Electrocardiographic Left Ventricular Hypertrophy with Blood Pressureand Blood Pressure Drop after Benazepril Treatment*

Systolic, mmHg Diastolic, mmHg

β ± s.e. P β ± s.e. PBaseline BP

Sokolow Lyon Index, mV 4.73 ± 0.66

The Journal of Applied Research • Vol. 4, No. 2, 2004 297

polymerase (Applied Biosystems,Foster City, Calif), and 1× reactionbuffer. The PCR reaction was carriedout using a DTC-225 thermocycler (MJResearch, Watertown, Mass) with an ini-tial denaturation at 94˚C for 10 minutes,followed by 14 touchdown cycles of 94˚Cfor 30 seconds, 61 to 54˚C for 30 seconds(decrease 0.5˚C each cycle), and 72˚C for30 seconds, then 30 cycles of 94˚C for 30seconds, 54˚C for 30 seconds, and 72˚Cfor 30 seconds, and finally 7 minutes ofextension at 72˚C. The PCR ampliconswere digested with 3 units of MspI (NewEngland Biolabs, Beverly, Mass) at 37˚Cfor 15 hours. Digested products wereseparated by electrophoresis on 2.5%agarose gels (FMC Bioproducts,Rockland, ME) and visualized withethidium bromide staining under ultravi-olet illumination.

Statistical AnalysisSAS 6.12 (SAS Institute, Inc.) and Splus2000 Professional (Mathsoft Inc.) wereused in data management and analysis.The Sokolow-Lyon Index was calculatedas the sum of |SV1| and |RV5|. First, aresidual plot of blood pressure (BP) or

blood pressure decrease after treatment(∆BP) vs. SLI was generated using thelowess smoothing function. Residualswere generated from a linear regressionmodel adjusted for age, age2, weight,weight2, height, height2, body mass index(BMI), gender, smoking status and alco-hol use. The plot of ∆BP was also con-trolled for baseline blood pressure.

Because the plots suggested a linearrelation, we performed regression analy-sis with multivariate linear additivemodels. In the model, BP or ∆BP wastreated as dependent variable, and SLIwas the primary independent variable,other covariates also entered the modelto adjust for potential bias. Gender is apotential effect modifier in this study, weinitially stratified the analysis by gender;however, no evidence of effect modifica-tion was detected, so the data were com-bined and we adjusted for gender inregression models. In examining thepotential synergetic effect of SLI andADRB3 genotype on the outcome (BPor ∆BP), we introduced the interactionterm (SLI × ADRB3) into the model.Also, we performed stratified analysis toexplore whether the relationship

Table 3. Interaction of Electrocardiographic Left Ventricular Hypertrophy and ADRB3 W64RGenotype on Blood Pressure and Blood Pressure Drop after Treatment*

Systolic, mmHg Diastolic, mmHg

ADRB3 β ± s.e. P β ± s.e. P

Baseline BP

Sokolow Lyon Index, mV 6.48 ± 0.98

between SLI and BP or ∆BP remainsconsistent in various genetic back-grounds. Because the allele frequencyof R64 is relatively low in our sample,we combined the W64R and R64R indi-viduals.

RESULTSA total of 1,199 hypertensive personsfrom Huoqiu County were initiallyrecruited for this study. Of these, 129failed to meet our enrollment criteria,102 did not complete the follow-uptreatment, and 44 were excluded fromthe study because major phenotypic

information was missing. The remaining924 subjects were included in the analy-sis of blood pressures. Two morepatients were dropped from the ADRB3analysis because of incomplete genotypeinformation. A clinical summary of thestudy cohort is presented in Table 1.All participants were diagnosed withessential hypertension and had used noanti-hypertensive therapy for at least 2weeks prior to the study; most subjects(86.0%) had never received medicationfor hypertension. Mean blood pressurebefore treatment was 149.4 mmHg sys-tolic and 93.4 mmHg diastolic. After

Vol. 4, No. 2, 2004 • The Journal of Applied Research298

Figure 1. Residual plot of baseline blood pressure (BP) and blood pressure drop after treatmentDBP) vs Sokolow-Lyon Index with lowess smoothing. Residuals were adjusted for age, age

2,weight, weight2, height, height2, BMI, smoking status, alcohol drinking, and baseline blood pres-sure (only in plotting of DBP). Lowess window was set as 2/3. Y-axis for DBP plot is the bloodpressure drop after treatment (BPbaseline-BPafter treatment), therefore larger value indicates better ther-apeutic effect.

treatment with benazepril for 15 days,blood pressure decreased to 9.9 mmHgand 5.2 mmHg in systolic and diastolicbranches respectively.

To examine the relationshipbetween SLI and blood pressure or itsdrop, a residual-residual smoothing plotwas first constructed (Figure 1, PanelsA-D). All four smoothing curves sug-gested linear patterns. SLI was positive-ly associated with baseline systolic anddiastolic BP (Panels A and B), and neg-atively associated with ∆BP for bothsystolic and diastolic pressures (PanelsC and D). The plots of BP and ∆BPagainst SLI quartiles also showed simi-lar linear trends (Figure 2).

ECG-diagnosed LVH was signifi-cantly associated with both BP and ∆BP

(Table 2). Each 1 mV increase in SLIwas associated with 4.73 mmHg(P

Vol. 4, No. 2, 2004 • The Journal of Applied Research300

ADRB3 wild type, the magnitude of theeffect associated with SLI increased,whereas among subjects with the mutantform of ADRB3, the associationbetween SLI and BP was attenuated anddid not reach statistical significance.Due to limited sample size and low fre-quency of ADRB3 R64 allele, the inter-action term (SLI × ADRB3) is notsignificant in the model. In addition, wefound the SLI was not associated withADRB3 genotype.

DISCUSSIONElectrocardiographically diagnosedLVH has long been considered a potentpredictor of hypertension and cardiovas-cular events.6,10 as well as a consequenceof elevated blood pressure andincreased afterload.9 In this study, weobserved a strong positive associationbetween SLI and resting BP in hyper-tensive subjects, and we found that sub-jects with a lower SLI had a morepronounced response to benazepriltreatment. We also found evidence ofan interaction between SLI and ADRB3genotype in the therapeutic response tobenazepril treatment: although SLI wassignificantly associated with both BPand ∆BP in subjects with the wild-typeADRB3, no association between SLIand either BP or ∆BP was observed inpersons with the mutant form ofADRB3.

Our findings are biologically mean-ingful. The inter-individual variation indrug response to antihypertensive med-ications is partially due to the complexnature of hypertension. For instance,the degree of LVH could reflect a partic-ular subtype of essential hyperten-sion.9,10,24 such that patients withdifferent SLI showed variant drugresponse. As a consequence of hyper-tension, LVH was correlated the bloodpressure in our study. Patients withLVH may have more severe or pro-longed hypertension, and their response

less response to ACEI treatment maydue to the geometry change in the car-diovascular system. ADRB3 R64 muta-tion modified the receptor function andis reported in association with insulinresistance, obesity and LVH, which areclosely related with hypertension statusand perhaps its subtypes.14-19

Our investigation possessed severalmajor advantages in avoiding potentialbias and improving power. (1) Beforethe administration of the benazeprilmonotherapy, no hypertensive subjectwas taking any other medication and thepotential confounding due to other drugeffects or drug-drug interaction wasexcluded. (2) All hypertensives werefree from other cardiovascular condi-tions such as coronary heart disease ordiabetes mellitus at the baseline. (3)The study population has settled forhundreds of years and is thought to behomogeneous genetically. (4) The resi-dential/occupational environmental anddietary habits were similar among ourstudy subjects, by this means, we con-trolled for potential socio-economic con-founders, which is difficult to measureand to quantify.

Managing essential hypertension is acomplicated task, because individualsmay respond differently to the sametype of drug.25 Our results have extend-ed new horizons in cardiovascular medi-cine by demonstrating the predictivevalue of electrocardiographic LVH inanti-hypertensive treatment effective-ness. The underlying mechanisms clearlydeserve further studies to resolve. Ifconfirmed independently by anotherinvestigation, our results could have sig-nificant clinical implications: the choiceof medicine and the therapeutic doserange can be predicted based on theECG indexes, which is widely and inex-pensively available. That will allow foroptimizing treatment effect at the indi-vidual level. Furthermore, there is con-siderable opportunity for future

The Journal of Applied Research • Vol. 4, No. 2, 2004 301

investigations to deepen our under-standing of hypertension subtypes andtheir responses to different therapeuticregimens.

REFERENCES1. Berkin KE, Ball SG. Essential hypertension:

the heart and hypertension. Heart.2001;86:467-475.

2. Rutherford S, Johnson MP, Curtain RP,Griffiths LR. Chromosome 17 and theinducible nitric oxide synthase gene in humanessential hypertension. Hum Genet.2001;109:408-415.

3. De Feo P, Torlone E, Perriello G, et al. Short-term metabolic effects of the ACE-inhibitorbenazepril in type 2 diabetes mellitus associ-ated with arterial hypertension. DiabeteMetab. 1992;18:283-288.

4. Wu Z, Bao X. Effects of benazepril on insulinresistance and glucose tolerance in uremia.Clin Nephrol. 1998;50:108-112.

5. Massie BM. Demographic considerations inthe selection of antihypertensive therapy. AmJ Cardiol. 1987;60:121I-126I.

6. Jern S. Assessment of left ventricular hyper-trophy in patients with essential hyperten-sion. Blood Press Suppl. 1997;2:16-23.

7. Tamama K, Kanda T, Osada M, Nagai R,Suzuki T, Kobayashi I. Detection of left ven-tricular enlargement by electrocariography. JMed. 1998;29:231-236.

8. Prisant LM. Hypertension images: electrocar-diographic left ventricular hypertrophy. J ClinHypertens. 2001;3:389-391, 398.

9. Balogun MO, Dunn FG. Left ventricularhypertrophy as a risk factor in hypertension.Afr J Med Med Sci. 1996;25:277-283.

10. Devereux RB, Roman MJ. Left ventricularhypertrophy in hypertension: stimuli, pat-terns, and consequences. Hypertens Res.1999;22:1-9.

11. Kahan T. The importance of left ventricularhypertrophy in human hypertension. JHypertens Suppl. 1998;16:S23-29.

12. Devereux RB, Okin PM, Roman MJ. Leftventricular hypertrophy as a surrogate end-point in hypertension. Clin Exp Hypertens.1999;21:583-593.

13. Strosberg AD. Structure and function of thebeta 3-adrenergic receptor. Annu RevPharmacol Toxicol. 1997;37:421-450.

14. Widen E, Lehto M, Kanninen T, Walston J,Shuldiner AR, Groop LC. Association of apolymorphism in the beta 3-adrenergic-recep-tor gene with features of the insulin resist-

ance syndrome in Finns. N Engl J Med.1995;333:348-351.

15. Clement K, Vaisse C, Manning BS, et al.Genetic variation in the beta 3-adrenergicreceptor and an increased capacity to gainweight in patients with morbid obesity. NEngl J Med. 1995;333:352-354.

16. Walston J, Lowe A, Silver K, et al. The beta3-adrenergic receptor in the obesity and dia-betes prone rhesus monkey is very similar tohuman and contains arginine at codon 64.Gene. 1997;188:207-213.

17. Kurabayashi T, Carey DG, Morrison NA. Thebeta 3-adrenergic receptor gene Trp64Argmutation is overrepresented in obese women.Effects on weight, BMI, abdominal fat, bloodpressure, and reproductive history in an eld-erly Australian population. Diabetes.1996;45:1358-1363.

18. Xiang K, Jia W, Lu H, et al. Effects ofTrp64Arg mutation in the beta 3-adrenergicreceptor gene on body fat, plasma glucoselevel, lipid profile, insulin secretion andaction in Chinese. Zhonghua Yi Xue YiChuan Xue Za Zhi. 1998;15:337-340.

19. Ringel J, Kreutz R, Distler A, Sharma AM.The Trp64Arg polymorphism of the beta3-adrenergic receptor gene is associated withhypertension in men with type 2 diabetesmellitus. Am J Hypertens. 2000;13:1027-1031.

20. Lind L, Andersson PE, Andren B, Hanni A,Lithell HO. Left ventricular hypertrophy inhypertension is associated with the insulinresistance metabolic syndrome. J Hypertens.1995;13:433-438.

21. Mureddu GF, Greco R, Rosato GF, et al.Relation of insulin resistance to left ventricu-lar hypertrophy and diastolic dysfunction inobesity. Int J Obes Relat Metab Disord.1998;22:363-368.

22. Watanabe K, Sekiya M, Tsuruoka T, FunadaJ, Kameoka H. Effect of insulin resistance onleft ventricular hypertrophy and dysfunctionin essential hypertension. J Hypertens.1999;17:1153-1160.

23. Buffone GJ, Darlington GJ. Isolation of DNAfrom biological specimens without extractionwith phenol. Clin Chem. 1985;31:164-165.

24. Devereux RB, Bella J, Boman K, et al.Echocardiographic left ventricular geometryin hypertensive patients with electrocardio-graphic left ventricular hypertrophy: TheLIFE Study. Blood Press. 2001;10:74-82.

25. Laragh JH, Lamport B, Sealey J, AldermanMH. Diagnosis ex juvantibus. Individualresponse patterns to drugs reveal hyperten-sion mechanisms and simplify treatment.Hypertension. 1988;12:223-226.