Author + information
- Received June 21, 2001
- Revision received January 3, 2002
- Accepted January 11, 2002
- Published online April 3, 2002.
- A.Mark Richards, MD, PhD†,* (, )
- M.Gary Nicholls, MD, FACC*,
- Richard W Troughton, PhD*,
- John G Lainchbury, MD†,
- John Elliott, MB, ChB, PhD†,
- Christopher Frampton, PhD*,
- Eric A Espiner, MD*,
- Ian G Crozier, MD*,
- Timothy G Yandle, PhD* and
- John Turner, MD‡
- ↵*Reprint requests and correspondence:
Dr. A. Mark Richards, Department of Medicine, Christchurch Hospital, Riccarton Avenue, P O Box 4345, Christchurch, New Zealand.
Objectives We sought to assess the relationship of antecedent hypertension to neurohormones, ventricular remodeling and clinical heart failure (HF) after myocardial infarction (MI).
Background Heart failure is a probable contributor to the increased mortality observed after MI in those with antecedent hypertension. Hence, neurohormonal activation, adverse ventricular remodeling and a higher incidence of clinical HF may be expected in this group. However, no previous report has documented serial postinfarction neurohumoral status, serial left ventricular imaging and clinical outcomes over prolonged follow-up in a broad spectrum of patients with and without antecedent hypertension.
Methods Inpatient events were documented in 1,093 consecutive patients (436 hypertensive and 657 normotensive) with acute MI. In 68% (282 hypertensive, 465 normotensive) serial neurohormonal sampling and radionuclide ventriculography were performed one to four days and three to five months after infarction. Clinical outcomes were recorded over a mean follow-up of two years.
Results Plasma neurohormones were significantly higher in hypertensives than in normotensives one to four days and three to five months after infarction. From similar initial values, left ventricular volumes increased significantly in hypertensives, compared with normotensives. Left ventricular ejection fraction rose significantly in normotensive but not hypertensive patients. Together with higher inpatient (8.1% vs. 4.4%, p < 0.002) and post-discharge mortality (9.5% vs. 5.5%, p = 0.043), hypertensive patients incurred more inpatient HF (33% vs. 24%, p < 0.001) and more late HF requiring readmission to hospital (12.4% vs. 5.5%, p < 0.001). Antecedent hypertension predicted late HF in patients >64 years of age with neurohormonal activation and early left ventricular dilation.
Conclusions Antecedent hypertension interacts with age, neurohumoral activation and early ventricular remodeling to confer greater risk of HF after MI.
Although antecedent hypertension is known to adversely affect mortality after acute myocardial infarction (MI) (1–9), whether (and in which patients) it also heightens the risk of developing heart failure (HF) is disputed (3,6,10–12). Furthermore, the potential interactions of antecedent hypertension with other predictors of postinfarction HF (including neurohormonal activation and left ventricular dimensions) are not well defined. Because hypertension induces structural changes within the left ventricle (13–17), we hypothesized that patients with hypertension would more often develop clinical HF and would exhibit greater activation of neurohormonal systems and more frequent adverse ventricular remodeling after acute MI than would normotensive individuals. The current report is the first to document serial postinfarction neurohormonal status and serial left ventricular imaging (radionuclide scanning) together with prolonged follow-up of clinical outcomes in a substantial series of patients with and without antecedent hypertension.
The study population was 1,093 patients admitted to the Christchurch Hospital Coronary Care Unit (CCU) with acute MI between November 1994 and June 1999. Acute MI was defined by the combination of typical symptoms together with ischemic changes in two or more ECG leads, and elevation of plasma creatine kinase (CK) to at least twice the upper limit of normal (400 U/l). Criteria for inclusion were age <80 years, absence of cardiogenic shock and survival for at least 24 h after onset of symptoms. Patients were categorized as having antecedent hypertension if this diagnosis was known by the patient to have been made by their family physician or after specialist referral, if the acute admission note indicated a history of hypertension and/or they were receiving antihypertensive medication.
Seven hundred and forty seven patients (68% of the group) gave written informed consent for additional tests and follow-up, according to a protocol approved by the Ethics Committee (Canterbury) of the Health Funding Authority of New Zealand. Blood samples for measurement of plasma atrial natriuretic peptide (ANP), brain natriuretic peptide (BNP), the amino terminal congeners of both ANP and BNP (N-ANP and N-BNP), cyclic guanosine monophosphate (cGMP), plasma catecholamines, adrenomedullin (ADM) and plasma endothelin-1 (ET-1) (18–24)were collected 24–96 h after the onset of symptoms through an indwelling intravenous cannula placed at least 30 min before sampling, with the patient resting quietly, semirecumbent. Samples were taken into chilled EDTA vacutainers, centrifuged within 20 min at 4°C and the plasma stored at −80°C before assay. Radionuclide ventriculography was performed within 24 h of blood sampling, with a General Electric 400 (Philadelphia, Pennsylvania) AC gamma camera interfaced to a General Electric 3000I computer system after standard in vivo technetium 99M red blood cell labeling. In survivors, both neurohumoral blood sampling and radionuclide scanning were repeated three to five months after MI.
Clinical events, including death, predischarge HF (presence of new symptoms of dyspnea and/or edema, with one or more of ventricular gallop rhythm, pulmonary crepitations, elevated venous pressure and/or radiologic evidence of left ventricular failure) and later readmissions for HF (similarly defined), were recorded for a mean follow-up period of two years. Patients were reviewed at 4 and 12 months after MI, then received questionnaires on clinical progress (with telephone follow-up) annually. Deaths were confirmed by death certificate and by consultation with medical staff responsible for inpatient care or with the family physician.
Values are expressed as mean ± standard deviation or standard error of the mean. Categorical clinical and demographic features and mean values of plasma hormones and nuclear ventriculography scan data were compared between normotensive and hypertensive groups by chi-square or independent ttests. Adverse event rates were compared (log rank test) by Kaplan-Meier survival analysis. Cox proportional hazards or multiple regression analyses were conducted to test the independent predictive power of antecedent hypertension and other putative indicators for predefined end points, including inpatient left ventricular failure and postdischarge cardiac failure requiring readmission to hospital. A p value <0.05 (two-tailed) was considered statistically significant.
Of 1,093 patients, 436 (40%) had antecedent hypertension. Hypertensive patients were, overall, slightly older; a greater proportion were female, had a higher body mass index, were more likely to have diabetes and/or renal dysfunction (plasma creatinine >0.11 mmol/l) and had more previously recognized dyslipidemia, compared with normotensive patients (Table 1). On the other hand, smoking was less prevalent in hypertensives than in normotensives. Rates of previous MI and HF were similar.
Peak CK levels (2,066 ± 1,724 vs. 2,199 ± 1,983 U/l; p = 0.265) and the proportion of patients suffering anterior MIs (42% and 39%; p = 0.336) were similar in hypertensives and normotensives, respectively. There was a trend towards more frequent primary ventricular fibrillation in hypertensives (8.2% vs. 5.8%; p = 0.151 NS). Thrombolytic therapy was administered in a slightly greater proportion of normotensives (61%) than of hypertensive patients (54%; p = 0.020), although in the subgroup (n = 747) undergoing more detailed testing and follow-up, this difference was not significant.
At discharge, rates of angiotensin-converting enzyme (ACE) inhibitor (53% vs. 36%, p < 0.001) and diuretic (27% vs. 16%, p < 0.001) use were significantly greater in hypertensive patients than in normotensive patients, whereas prescription of beta-blockers, antilipid agents and aspirin did not differ.
Plasma norepinephrine, the cardiac natriuretic peptides, their N-terminal propeptides and plasma cGMP levels were significantly higher in the hypertensive group one to four days after MI and/or three to five months after MI (Table 2), with these differences becoming statistically more robust with the passage of time. In contrast, plasma epinephrine, ADM and ET-1 levels did not differ between groups at either time point.
Radionuclide ventriculography indicated similar or smaller left ventricular systolic (85 ± 2.5 vs. 89 ± 2.3 ml; p = 0.28) and diastolic (152 ± 2.9 vs. 160 ± 2.6 ml; p = 0.041) volumes and mean ejection fractions (46.8 ± 0.8 vs. 46.4 ± 0.6%; p = 0.656 NS) in hypertensives compared to normotensives, one to four days after MI. Left ventricular volumes increased in hypertensive patients, compared with normotensive patients, over the three- to five-month follow-up period (p < 0.05 to p < 0.001). Ejection fraction rose in normotensive but not hypertensive patients (p < 0.01; Fig. 1).
Multiple regression analysis for independent prediction of increases in left ventricular systolic volume after MI indicated the following. Of 13 demographic, clinical, pharmaceutical and neurohormonal indicators (including age, gender, peak CK, previous MI, previous HF, diabetes, renal impairment, dyslipidemia, ACE inhibitor use at discharge, beta-blocker use at discharge, diuretic use at discharge, plasma BNP and hypertension), antecedent hypertension was the single most powerful independent predictor (p < 0.001), with age (p = 0.003) and plasma BNP (p = 0.03) also remaining significant. When all other neurohormonal variables were rotated through this model in place of BNP, none attained statistical significance.
Mortality and HF
Hypertensive patients, compared with normotensive patients, had a higher inpatient mortality (8.1% vs. 4.4%; p < 0.002) and postdischarge mortality (9.5% vs. 5.5%; p = 0.043; Fig. 2). Hypertensives also had a greater chance of developing acute left ventricular failure (33% vs. 24%; p < 0.001) or later HF requiring readmission to hospital (12.4% vs. 5.5%; p < 0.001; Fig. 2). Multivariate testing indicated antecedent hypertension, age, previous MI, diabetes, renal impairment and peak CK levels were all independently predictive of inpatient LVF (p < 0.05 to p < 0.001).
Multivariate analysis incorporating seven putative predictors (age, gender, peak CK, left ventricular ejection fraction, history of diabetes, previous MI and hypertension) indicated antecedent hypertension, age, left ventricular ejection fraction and previous MI were all significant independent predictors of postdischarge cardiac failure (p ≤ 0.002 for all). These variables remained significant when neurohumoral factors were separately rotated through this multivariate model (as an eighth variable) and early plasma concentrations of BNP (p < 0.001), N-BNP (p < 0.01), ANP (p < 0.001), N-ANP (p < 0.05) and endothelin (p < 0.05) were all also significantly and independently predictive of late HF.
Significant interactions were seen between a diagnosis of antecedent hypertension and the following variables (for which interaction terms were separately rotated through the multivariate model in place of the simple “hypertension” term): age; peak CK; left ventricular end systolic volume; left ventricular end diastolic volume; and initial plasma levels of BNP, N-BNP, N-ANP, endothelin, ADM (ANP, cGMP, norepinephrine and epinephrine), for the independent prediction of post-discharge HF. The concurrence of antecedent hypertension with above-median levels of age, peak CK, any of the above listed neurohormones or of left ventricular volumes was independently associated with significantly (p < 0.01 for all interactions) increased risk of HF. Particularly, striking interactions modifying the relationship of antecedent hypertension to risk of postdischarge HF were observed for age, plasma N-BNP and left ventricular end systolic volume (all p < 0.0001) as illustrated in detail in Figure 3.
Cumulative admissions for HF in younger (below median, i.e., <64 years of age) hypertensive patients (4.5%) exceeded those in younger normotensive patients (3.3%) by only 1.2% (p = NS). Corresponding event rates were 19.7% and 7.9%, respectively, in older patients (i.e., an absolute difference of 11.8%, p < 0.001; 9.8 times the difference observed in younger patients). Corresponding event rates and the intergroup differences for patients with N-BNP below versus above the group median (120 pmol/l) were 4.0% versus 2.0% (difference 2.0%, p = NS) and 20.5% versus 11.1% (difference 9.4%, p < 0.001; 4.7 times the difference observed in the absence of elevated N-BNP). Finally, for left ventricular end-systolic volume below versus above the median (78 ml), corresponding event rates and intergroup differences were, respectively, 5.8% versus 1.9% (difference 3.9%, p = NS) and 20.3% versus 8.1% (difference 12.2%, p < 0.001; 3.1-fold the difference in the subgroups with smaller systolic volumes).
Hypertension is a well-established risk factor for coronary artery disease and acute MI (25,26). Effects of chronic hypertension, which may predispose patients to an adverse effect on outcomes after acute MI, include increased coronary vascular resistance, decreased coronary reserve, pre-existing muscle fiber hyperplasia and augmented interstitial collagen together with possible alterations in the responsiveness of the coronary vasculature to vasoactive transmitters (13–17). The weight of evidence indicates an adverse effect of antecedent hypertension on survival after MI (1–9).
Hypertension is also a common substrate for HF (27–29), which has a high mortality (28). In those with antecedent hypertension, increased propensity to HF after MI may contribute to this group’s increased postinfarction mortality. However, reports of the relative frequency of acute and/or chronic postinfarction HF in hypertensive compared with normotensive patients, are scant and conflicting (3,6,10–12).
We assessed the importance of antecedent hypertension in a large cohort recruited from a single center over a five-year period. Consecutive patients (except for study exclusions and declined consent) were studied. Follow-up averaged two years. To our knowledge, this is the first report from a substantial consecutive series of patients with MI (including a broad spectrum of severity) that compares outcomes in normotensive and hypertensive patients and incorporates serial left ventricular imaging and serial neurohormonal profiling, in addition to documenting clinical outcomes.
We found that patients with hypertension exhibited greater activation of neurohormones, both in the early postinfarction period and several months later. Such neurohumoral activation has been shown to have prognostic significance after MI (30–32). This distinction between groups was observed for the cardiac natriuretic peptides and plasma norepinephrine. This implies greater cardiac hemodynamic stress for a given infarct size in hypertensive patients. Plasma levels of the cardiac peptides reflect intracardiac pressures (33,34). Ventricular filling pressures were, therefore, presumably greater in the hypertensive group. Diastolic function also influences plasma cardiac peptide levels (35)and is likely to have been more impaired in hypertensive ventricles secondary to hypertrophic and interstitial changes attendant upon chronic hypertension.
In association with neurohumoral activation, significant left ventricular dilation occurred in the hypertensive group, with mean increases in systolic and diastolic volume clearly greater in hypertensive than in normotensive subjects. Both antecedent hypertension and plasma BNP were significant independent predictors of increasing left ventricular systolic volume.
Mortality and HF
Most importantly, our data indicate that patients with antecedent hypertension suffer increased acute and chronic mortality and cardiac failure after MI. This finding is consistent with evidence indicating that 90% of patients who eventually present with overt cardiac failure have an antecedent history of hypertension (27). The association of antecedent hypertension with adverse outcomes is amplified by age >64 years, more marked neurohumoral activation (N-BNP >120 pmol/l) and more pronounced left ventricular dilation (LVESV >78 ml). In fact, in the absence of these features, hypertension carried no overall adverse prognostic significance (Fig. 3).
Differences between normotensive and hypertensive groups in age, gender, body mass index, smoking and diabetes confirmed multiple previous reports (3,5,6). However, when adjustment (multivariate analysis) was made for these baseline differences, antecedent hypertension continued to independently predispose patients to HF. Notably, peak CK levels, the frequency of anterior MI and the early postinfarct left ventricular ejection fraction and dimensions did not indicate greater initial infarct size in the hypertensive group. Furthermore, greater neurohormonal activity, mortality and morbidity occurred in those with antecedent hypertension, despite significantly greater prescription of ACE inhibitors and diuretics than in normotensive subjects (53% vs. 36% and 27% vs. 16%, respectively; p < 0.001 for both comparisons). Therefore, the hypertensive heart is more vulnerable to adverse left ventricular remodeling and progression to frank congestive cardiac failure for a given infarct size. This may reflect the mechanisms as discussed previously (13–17).
Limitations of this report include inability to comment on outcomes in the age >80 years group or on any non-Caucasian patient populations. Also, we did not have detailed serial pre-infarction blood pressure recordings. Furthermore, the effects of MI, bed rest and postinfarct drug therapy probably obscured new diagnosis of hitherto unrecognized hypertension. These factors and prospective epidemiologic evidence (1)suggest that underestimation of the true prevalence of antecedent hypertension in our study population is likely. However, our observed prevalence of 40% matches other recent reports (3,4,9)and our data indicate that antecedent hypertension as defined in the current report is of clinical importance.
In summary, antecedent hypertension interacts with patient age, neurohormonal status and adverse ventricular remodeling to confer increased risk of progression to congestive cardiac failure after MI.
The authors gratefully acknowledge the assistance of Mrs. Rona Buttimore (Senior Research Nurse), Ms. Justine Miller, Ms. Leanne Liggett and Mrs. Rose Richards (Research Assistants), consultant, nursing and technical staff of the Departments of Cardiology, Nuclear Medicine and Endocrinology. Secretarial assistance was provided by Mrs. Barbara Griffin.
☆ Supported by grants from the National Heart Foundation of New Zealand and the Health Research Council of New Zealand. Dr. Troughton holds a National Heart Foundation (New Zealand) Research Fellowship. Dr. Richards holds the National Heart Foundation (New Zealand) Chair of Cardiovascular Studies.
- angiotensin-converting enzyme
- atrial natriuretic peptide
- brain natriuretic peptide
- cyclic guanosine monophosphate
- creatine kinase
- endothelin 1
- heart failure
- myocardial infarction
- aminoterminal atrial natriuretic peptide
- aminoterminal brain natriuretic peptide
- Received June 21, 2001.
- Revision received January 3, 2002.
- Accepted January 11, 2002.
- American College of Cardiology Foundation
- Wong N.D,
- Cupples L.A,
- Ostfeld A.M,
- et al.
- Volpi A,
- De Vita C,
- Franzosi M.G,
- et al.
- Haider A.W,
- Chen L,
- Larson M.G,
- et al.
- Gustafsson F,
- Køber L,
- Torp-Pedersen C,
- et al.
- Herlitz J,
- Bång A,
- Karlson B.W
- Rutherford J.D,
- Pfeffer M.A,
- Moyé L.A,
- et al.
- Spencer F.A,
- Meyer T.E,
- Goldberg R.J,
- et al.
- Schwartzkopff B,
- Strauer B.E
- Yandle T,
- Espiner E,
- Nicholls M,
- et al.
- Steiner A,
- Parker C,
- Kipnis D
- Lewis L.K,
- Smith M.W,
- Yandle T.G,
- et al.
- Hall C,
- Rouleau J.L,
- Moyé L,
- et al.
- Omland T,
- Aakvaag A,
- Bonarjee V.V.S,
- et al.
- Bates E.R,
- Shenker Y,
- Grekin R.J
- Richards A.M,
- Crozier I.G,
- Yandle T.G,
- et al.