Patients with isolated systolic hypertension (ISH)

Transcription

1 AJH 2001; 14: Association of Increased Pulse Pressure With the Development of Heart Failure in SHEP John B. Kostis, Janet Lawrence-Nelson, Rajiv Ranjan, Alan C. Wilson, William J. Kostis, and Clifton R. Lacy, for the Systolic Hypertension in the Elderly (SHEP) Cooperative Research Group The aim of this study was to assess the relationship between pulse pressure (PP) and the occurrence of heart failure (HF) in older persons with isolated systolic hypertension. Data from a prospective, multicenter, randomized, double-blind, placebo-controlled clinical trial were analyzed. A total of 4736 persons aged 60 years with systolic blood pressure (SBP) between 160 and 219 mm Hg and diastolic blood pressure (DBP) 90 mm Hg who participated in the Systolic Hypertension in the Elderly Program (SHEP) were studied. The main outcome measures were fatal and nonfatal HF. During 4.5 years average follow-up, fatal or nonfatal HF occurred in 160 of 4736 patients. The SBP, PP, and mean arterial pressure (MAP) were strong predictors of the development of HF (P.0002). Cox proportional hazards regression using timedependent covariates and controlling for MAP indicated that HF was inversely related to DBP (P 0.002) and was directly related to pulse pressure (P 0.002). Data were similar when patients who developed myocardial infarction during follow up were excluded. These data indicate that, in older persons with isolated systolic hypertension, high pulse pressure is associated with increased risk of heart failure independently of MAP and of the occurrence of acute myocardial infarction during follow-up. Am J Hypertens 2001;14: American Journal of Hypertension, Ltd. Key Words: Pulse pressure, heart failure, isolated systolic hypertension. Patients with isolated systolic hypertension (ISH) have increased pulse pressure (PP) and greater risk for cardiovascular events including stroke, myocardial infarction, and heart failure (HF). 1 3 Large-scale controlled clinical trials of antihypertensive therapy in older patients with isolated systolic hypertension have demonstrated a decrease in clinical events including HF. 4 6 Increased PP has been associated with greater likelihood of myocardial infarction, stroke, and cardiovascular mortality This has been attributed to the occurrence of ischemic events caused by low diastolic pressure in patients with coronary artery disease. 7,12 This mechanism is unlikely to be the primary determinant of HF in the absence of intervening myocardial infarction. Chae et al reported an increased risk of HF in subjects with increased PP 13 ; in their study, mean blood pressure (BP) was 137/75 mm Hg and mean PP 62 mm Hg. This report presents observations on the relationship of PP to the development of HF among 4736 men and women with ISH (mean BP 170/77 mm Hg and mean PP 93 mm Hg) who participated in the Systolic Hypertension in the Elderly program (SHEP). 4,6 Materials and Methods The SHEP trial was a placebo-controlled, double-blind, randomized, multicenter clinical trial testing the efficacy of diuretic-based stepped-care antihypertensive drug treatment of ISH in persons aged 60 years. 4,6 The major inclusion criterion was the presence of ISH, defined as systolic blood pressure (SBP) of 160 to 219 mm Hg with a diastolic blood pressure (DBP) 90 mm Hg. Patients with recent myocardial infarction (MI), coronary bypass surgery, dementia, stroke with residual paralysis, insulindependent diabetes mellitus, atrial fibrillation or flutter, arterioventricular block, multiform premature ventricular Received June 19, Accepted January 17, From the University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School (JBK, JL-N, RR, ACW, CRL), Department of Medicine, New Brunswick, New Jersey; Center for Disease Management and Clinical Outcomes and School of Electrical Engineering (WJK), Cornell University, Ithaca, New York. This study was supported by contracts with the National Heart, Lung, and Blood Institute and the National Institute on Aging. Drugs were supplied by the Lemmon Co, Sellersville, PA; Wyeth Laboratories/ Ayerst Laboratories, A.H. Robins Co, Richmond, VA; and Stuart Pharmaceuticals, Wilmington, DE. Address correspondence and reprint requests to John B. Kostis, MD, Department of Medicine, UMDNJ-Robert Wood Johnson Medical School, One Robert Wood Johnson PI., PO Box 19, New Brunswick, NJ ; kostis@umdnj.edu /01/$ by the American Journal of Hypertension, Ltd. Downloaded from PII S (01) Published by Elsevier Science Inc.

2 AJH August 2001 VOL. 14, NO. 8, PART 1 PULSE PRESSURE AND HEART FAILURE IN SHEP 799 contractions, bradycardia 50 beats/min on the 12-lead electrocardiogram (ECG), alcohol abuse, or need for diuretic therapy were excluded. Controlled heart failure was not an exclusion criterion. Although the primary endpoint of the study was total stroke, HF was a prespecified secondary endpoint. The main results of SHEP including findings on HF; the criteria for enrollment have been described in detail elsewhere. 4,6 Patients were randomized to receive either placebo or active stepped-care therapy consisting of chlorthalidone 12.5 mg daily, which was stepped up to 25 mg if needed to reach target blood pressure, with atenolol 25 mg stepped up to 50 mg if needed as second-step drug. If atenolol was contraindicated, 0.05 to 0.1 mg reserpine or matching placebo daily was substituted. Target BP was a decrease in SBP by 21 mm Hg and to a level of 159 mm Hg. Information related to BP, interval history, and study endpoints was collected by clinic staff. When MI or HF was suspected, ECG, cardiac enzymes, chest roentgenogram reports, and other clinical information was obtained. For hospitalizations and nursing home admissions, discharge and/or admission records were obtained. Death certificates and autopsy reports were collected when available. A coding panel of three physicians including at least one cardiologist, all blinded to randomization allocation, confirmed the study endpoints. The diagnosis of HF was made by the committee when one of the following symptoms was present: dyspnea at rest, orthopnea, paroxysmal nocturnal dyspnea, or New York Heart Association class III. In addition, the presence of one or more of the following signs or laboratory findings was required: rales, ankle edema ( 2 ), tachycardia of 120 beats/min, cardiomegaly by chest x-ray, chest x-ray characteristic of congestive failure, S3 gallop, or jugular venous distention. The diagnosis of HF was not made in the presence of severe pulmonary disease manifested by chronic obstructive pulmonary disease (including positive chronic bronchitis questionnaire, plus smoking history of 10 packyears and x-ray confirmation), pneumonia, or other severe lung disease documented by x-ray or other tests. Fatal and nonfatal HF occurring after randomization (n 160) were considered in this analysis. Using the 2 test for categorical variables and analysis of variance for continuous variables, we examined baseline PP groups by tertiles (70 to 86, 87 to 96, and 96 mm Hg) for significant differences. Mean arterial pressure (MAP) was calculated as (2 DBP 1 SBP)/3. The relationships between BP during follow-up and the risk of HF were studied using univariate and multivariate Cox regression 11 models using SBP, DBP, PP, and MAP as time-dependent covariates. Relative risks (RR) and their 95% confidence intervals (CI) were calculated based on the entire duration of follow-up controlling for age, sex, diabetes, and treatment group. These factors were related to the occurrence of heart failure during follow-up by univariate Cox analysis (P.05). The relationship between PP and HF was also examined in each tertile of DBP and each tertile of DBP using univariate Cox regression. Three patients with missing baseline PP were excluded from the analysis leaving 157 patients. During follow-up, three patients in the placebo group and four patients in the active treatment group experienced MI and then developed HF. Results The recruitment, randomization, and baseline characteristics of SHEP participants, as well as antihypertensive treatment during follow-up and its effects on SBP and DBP, adverse events, fatal and nonfatal stroke, heart failure, and morbidity and mortality from cardiovascular causes have been reported elsewhere. 4,6 A total of 4736 persons aged 60 years were randomized to treatment with active stepped-care therapy (n 2365) or placebo (n 2371). At baseline the average BP was 170.3/77.0 mm Hg, PP mm Hg, and MAP mm Hg. Sixteen patients (0.3%) had a history of HF and 492 participants (10.5%) had evidence of prior MI by history or ECG. Approximately 90% of participants randomized to the active treatment group remained on active antihypertensive medication during the study, either by protocol or by prescription of open-label drug. The percentage of patients randomized to placebo who received active medication increased from 13.1% in year 1 to 32.7% in year 3 and to 44.4% in year 5. 4 During follow-up, BP in the active treatment group averaged 143/68 mm Hg, PP mm Hg, and MAP mm Hg, compared with 155/72 mm Hg, PP mm Hg, and MAP mm Hg in the placebo group. Baseline characteristics and occurrence of HF at follow-up in patients categorized by tertiles of PP are shown in Table 1. Participants with higher PP were more likely to be older, to be women and to have diabetes and carotid bruits. They were less likely to be smokers or drinkers and had lower BMI. During 4.5 years average follow-up, fatal or nonfatal HF occurred in 160 participants. In three, the PP value is missing. Among both treatment groups combined, fatal or non-fatal HF was more likely to occur among participants with higher baseline PP (PP 70 to 86 mm Hg, 2.6%; PP 87 to 96, 3.0%; and 96 mm Hg, 4.4%; RR for 96 mm Hg v 96 mm Hg, RR 1.59, 95% CI 1.15 to 2.19) (Table 2). Baseline PP was associated with the development of HF during follow-up in the total cohort (P.009), in the total cohort after exclusion of patients who developed MI during follow-up (P.01), in the placebo group (P.041), and in the treatment group (P 0.011) (Table 2). By univariate Cox regression, HF occurred more frequently in men (RR 1.47, 95% CI 1.08 to 2.01); in older participants (RR for 80 years versus other ages 1.08, 95% CI 1.06 to 1.10); in those with higher baseline SBP (RR/per mm Hg 1.03, 95% CI 1.02 to 1.03) or higher PP (RR per mm Hg 1.02, CI 1.01 to 1.03), higher baseline MAP (RR per mm Hg 1.03, CI 1.01 to 1.04); and in those with diabetes (RR 2.10, 95% CI 1.38 to 3.18) (Table 3). Pulse pressure had

3 800 PULSE PRESSURE AND HEART FAILURE IN SHEP AJH August 2001 VOL. 14, NO. 8, PART 1 Table 1. pressure Baseline characteristics in the systolic hypertension in the elderly program of tertiles or pulse Characteristic N mm Hg mm Hg PP > 96 mm Hg Total P Value N Age, years, mean SD Women (%) African American (%) Diabetes (%) History of MI (%) BMI, kg/m Education high school (%) Alcohol 1 drink/week (%) Uric acid mg/dl, mean SD Smoking (%) On active therapy (%) Carotid bruits (%) SBP, mean SD DBP, mean SD MAP, mean SD pulse pressure; MI myocardial infarction; BMI body mass index; SBP systolic blood pressure; DBP diastolic blood pressure; MAP mean arterial pressure. a significant relationship with HF at different levels of DBP and MAP; it was associated with the occurrence of HF in all tertiles of DBP (RR per mm Hg, to 1.025, P.005), as well as in all tertiles of MAP (RR per mm Hg, to 1.032, P.001) (Table 4). In the Cox model, the BP variables were time dependent. Adjusted Cox proportional hazards regression using time-dependent covariates (SBP, MAP, DBP, PP) separately and in pairs indicated that HF was directly related to SBP, MAP, and PP, but not to DBP (Table 5). Data were similar when seven patients who developed nonfatal MI during follow-up were excluded (not shown). In pairwise analyses, the relationship of DBP to HF was positive after accounting for PP and was negative after accounting for MAP. Discussion The observation of higher risk of heart failure among SHEP participants with higher PP after adjusting for MAP (either by Cox regression or by subsetting into PP tertiles) is consistent with the observations of Chae et al in a lower BP cohort and with previous reports from SHEP on other cardiovascular endpoints such as stroke and MI. 10,11,13 Although the decreased DBP linked to high PP may trigger thrombotic events such as stroke and MI, it cannot fully explain the relationship between PP and HF. Increased afterload is considered to be the major hemodynamic cause of left ventricular failure in ISH without an intervening MI. 15 The increased risk of HF in patients with high PP can be explained by the association of decreased arterial compliance with both PP and increased afterload. Decreased arterial compliance causes both higher SBP and lower DBP, resulting in higher PP as well as higher impedance faced by the contracting ventricle. Reduced compliance also causes the reflected pressure wave to arrive early to the aortic root thus increasing left ventricular work In addition, high PP has been associated with LVH and impaired diastolic left ventricular relaxation that could present as diastolic HF, a condition prevalent in elderly patients. 19 Decreased diastolic relaxation may also be a manifestation of a generalized loss of elasticity with aging. 20,21 This stiffening may affect the Table 2. Occurrence of heart failure during follow-up by tertiles of baseline pulse pressure in the systolic hypertension in the elderly program N mm Hg mm Hg PP > 96 mm Hg 2 P for Trend RR (95% CI)* Total Events (2.6%) 49 (3.0%) 69 (4.4%) ( ) Placebo (3.4%) 35 (4.2%) 44 (5.6%) (1 2.22) Treatment (1.9%) 14 (1.7%) 25 (3.2%) ( ) CI confidence interval; other abbreviation as in Table 1. * Risk comparing the highest tertile to the lower two tertiles. Three missing values (one in placebo and two in treatment group).

4 AJH August 2001 VOL. 14, NO. 8, PART 1 PULSE PRESSURE AND HEART FAILURE IN SHEP 801 Table 3. Univariate relationships (Cox regression) of the occurrence of heart failure in SHEP Univariate RR (95% CI) P Value Pulse pressure* (mm Hg) 1.02 ( ).0001 SBP* (mm Hg) 1.03 ( ).0001 DBP* (mm Hg) 1.00 ( ).624 Mean Arterial Pressure* (mm Hg) 1.03 ( ).0002 Age (y) 1.08 ( ).0001 Treatment/Placebo 0.50 ( ).0001 Diabetes (1 yes; 0 no) 2.10 ( ).0002 Sex (1 M; 0 F) 1.47 ( ).014 Smoking (1 yes; 0 no) 1.33 ( ).194 Body mass index (kg/m 2 ) 1.01 ( ).413 Race (1 white; 0 not white) 0.83 ( ).316 Carotid bruit (1 yes; 0 no) 2.91 ( ).0001 Alcohol (1 yes; 0 no) 0.70 ( ).054 SHEP Systolic Hypertension in the Elderly Program; other abbreviations as in Tables 1 and 2. * Time dependent variables using follow-up values. large arteries and the left ventricle independently, resulting in both decreased arterial compliance and decreased left ventricular relaxation and resulting in turn in HF. It is possible that genetic and other factors such as diabetes mellitus and smoking affect the rate of development of these changes with age in the left ventricle and the aorta in a similar fashion. In that case, increased PP may be a marker of both low arterial compliance and decreased left ventricular relaxation It is also possible that decreased coronary blood flow due to low DBP may impair left ventricular function and cause HF. 26 However, the absence of ischemic symptoms and the persistence of the relationship of PP with HF after excluding patients with intervening MI suggest that this is not the primary mechanism. When each BP variable was entered alone in the multivariate models, a strong relationship of PP and SBP with HF was observed. The effect of DBP when entered in the model as the only BP variable was small and not statistically significant. When two BP variables were entered together in the model, higher DBP was associated with lower rate of HF after accounting for the effect of MAP. On the contrary, after accounting for PP, higher DBP was associated with higher rate of HF. These observations underscore the importance of decreased arterial compliance in increasing afterload and SBP. In the presence of a given mean pressure, a higher DBP implies lower SBP (explaining the negative relationship of DBP to HF), whereas in the presence of a given PP, a higher DBP implies higher SBP (explaining the positive relationship of DBP to HF). Table 4. Univariate Cox regression parameters obtained with pulse pressure as explanatory factor: stratified by tertiles of baseline MAP and DBP Cox Regression with PP as Time- Tertiles of Baseline DBP Dependent Factor mm Hg mm Hg > 80 mm Hg HF events* coefficient RR (95% CI) P Value Tertiles of Baseline MAP mm Hg mm Hg mm Hg HF events* coefficient RR (95% CI) P Value pulse pressure (PP units are mm Hg); HF heart failure; RR relative risk; other abbreviations as in Tables 1 3. * HF events do not total 160 because of missing values for blood pressure during follow-up.

5 802 PULSE PRESSURE AND HEART FAILURE IN SHEP AJH August 2001 VOL. 14, NO. 8, PART 1 Table 5. Occurrence of heart failure in SHEP modeled by adjusted Cox regression using blood pressure components as time dependent covariates singly and in pairs -Coefficient RR (95% CI) P Value Single component Pulse Pressure ( ).0001 SBP (mm Hg) ( ).0001 DBP (mm Hg) ( ).242 MAP (mm Hg) ( ).002 Pairs of components PP and ( ).002 MAP ( ).027 PP and ( ).0001 DBP ( ).027 SBP and ( ).0001 DBP ( ).625 SBP and ( ).002 MAP ( ).625 DBP and ( ).002 MAP ( ).0001 Abbreviations as in Tables 1 4. Measurement units for all blood pressure were mm Hg. Controlling for age, sex, treatment group, and diabetes. Regardless of mechanism, it appears that increased PP in patients with ISH is associated with higher risk of HF. Whether different classes of pharmacotherapy will have different effects on event reduction according to differential effect on lowering PP remains speculative. References 1. Smulyan H, Safar ME: The diastolic blood pressure in systolic hypertension. Ann Intern Med 2000;132: Gubner RS: Systolic hypertension: a pathogenetic entity: significance and therapeutic considerations.am J Cardiol 1962;9: Colandrea MA, Friedman GD, Nichaman MZ, Lynd CN: Systolic hypertension in the elderly. An epidemiologic assessment. Circulation 1970;41: SHEP Cooperative Research Group: Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension. Final results of the Systolic Hypertension in the Elderly program (SHEP). JAMA 1991;265: Staessen JA, Fagard R, Thijs L, Celis H, Arabidze GG, Birkenhager WH, Bulpitt CJ, de Leeuw PW, Dollery CT, Fletcher AE, Forette F, Leonetti G, Nachev C, O Brien ET, Rosenfeld J, Rodicio JL, Tuomilehto J, Zanchetti A: Randomized double-blind comparison of placebo and active treatment for older patients with isolated systolic hypertension. The Systolic Hypertension in Europe (Syst-Eur) Trial Investigators. Lancet 1997;350: Kostis JB, Davis BR, Cutler J, Grimm RH Jr, Berge KG, Cohen JD, Lacy CR, Perry HM Jr, Blaufox MD, Wassertheil-Smoller S, Black HR, Schron E, Berkson DM, Curb JD, Smith WM, McDonald R, Applegate WB, for the SHEP Cooperative Research Group: Prevention of heart failure by antihypertensive drug treatment in older persons with isolated systolic hypertension. JAMA 1997;278: Madhavan S, Ooi WL, Cohen H, Alderman MH: Relation of pulse pressure and blood pressure reduction to the incidence of myocardial infarction. Hypertension 1994;23: Benetos A, Safar M, Rudnichi A, Smulyan H, Richard JL, Ducimetieere P, Guize L: Pulse pressure: a predictor of long-term cardiovascular mortality in a French male population. Hypertension 1997;30: Mitchell GF, Moye LA, Braunwald E, Rouleau JL, Bernstein V, Geltman EM, Flaker GC, Pfeffer MA, for the SAVE investigators: Sphygmomanometrically-determined pulse pressure is a powerful independent predictor of recurrent events after myocardial infarction in patients with impaired left ventricular function. Circulation 1997; 96: Domanski MJ, Davis BR, Pfeffer MA, Kastantin M, Mitchell GF: Isolated systolic hypertension: prognostic information provided by pulse pressure. Hypertension 1999;34: Somes GW, Pahor M, Shorr RI, Cushman WC, Applegate WB: The role of diastolic blood pressure when treating isolated systolic hypertension. Arch Intern Med 1999;159: Franklin SS, Khan SA, Wong ND, Larson MG, Levy D: Is pulse pressure useful in predicting risk for coronary heart disease? The Framingham Heart Study. Circulation 1999;100: Chae CU, Pfeffer MA, Glynn RJ, Mitchell GF, Taylor JO, Hennekens CH: Increased pulse pressure and risk of heart failure in the elderly. JAMA 1999;281: Cox DR: Regression models and life-tables. J R Stat Soc 1972;34: Nichols WW, O Rourke MF, Avolio AP, Yaginuma T, Murgo JP, Pepine CJ, Conti CR: Effects of age on ventricular-vascular coupling. Am J Cardiol 1985;55: Safar ME, Laurent S, Asmar RE, Safavian A, London GM: Systolic hypertension in patients with arteriosclerosis obliterans of the lower limbs. Angiology 1987;38: Cohn JN, Finkelstein SM: Abnormalities of vascular compliance in hypertension, aging and heart failure. J Hypertens 1992;10(suppl 6):S61 S O Rourke MF, Brunner HR: Introduction to arterial compliance and function. J Hypertens 1992;10(suppl 6):S3 S Bouthier JD, De Luca N, Safar ME, Simon AC: Cardiac hypertrophy and arterial distensibility in essential hypertension. Am Heart J 1985;109: Klein AL, Burstow DJ, Tajik AJ, Zachariah PK, Bailey KR, Seward JB: Effects of age on left ventricular dimensions and filling dynamics in 117 normal persons. Mayo Clin Proc 1994;69: Schulman SP, Lakatta EG, Fleg JL, Lakatta L, Becker LC, Gerstenblith G: Age-related decline in left ventricular filling at rest and exercise. Am J Physiol 1992;263:H1932 H1938.

6 AJH August 2001 VOL. 14, NO. 8, PART 1 PULSE PRESSURE AND HEART FAILURE IN SHEP Benetos A, Gautier S, Ricard S, Topouchain J, Asmar R, Poirier O, Larosa E, Guize L, Safar M, Soubrier F, Cambien F: Influence of angiotensin-converting enzyme and angiotensin II type 1 receptor gene polymorphisms on aortic stiffness in normotensive and hypertensive patients. Circulation 1996;94: McVeigh GE: Arterial compliance in hypertension and diabetes mellitus. Am J Nephrol 1996;16: Kool MJ, Hoeks AP, Struijker Boudier HA, Reneman RS, Van Bortel LM: Short- and long-term effects of smoking on arterial wall properties in habitual smokers. J Am Coll Cardiol 1993;22: Blacher J, Staessen JA, Girerd X, Gasowski J, Thijs L, Liu L, Wang JG, Fagard RH, Safar ME: Pulse pressure not mean pressure determines cardiovascular risk in older hypertensive patients. Arch Intern Med 2000;160: Watanabe H, Ohtsuka S, Kakihana M, Sugishita Y: Coronary circulation in dogs with an experimental decrease in aortic compliance. J Am Coll Cardiol 1993;21: