Prognostic Value of Invasive Hemodynamic Measurements at Rest and During Exercise in Hypertensive Men
In 1994, we ascertained the outcome of 143 hypertensive men in whom invasive hemodynamic measurements were performed at rest and during graded bicycle exercise during the period 1972-1982 to assess (1) which of the hemodynamic components of blood pressure is associated with the incidence of cardiovascular events and total mortality, and (2) whether the hemodynamic response to dynamic exercise adds prognostic precision to the data at rest. During 2186 patient years of follow-up, 38 patients suffered at least one fatal or nonfatal cardiovascular event and 17 patients died. Cox regression analysis showed that systolic pressure and systemic vascular resistance measured at rest, during submaximal exercise (50 W), and at peak effort were significant (P<.01) predictors of the age-adjusted incidence of cardiovascular events and total mortality. However, exercise blood pressure did not significantly predict the incidence of cardiovascular events over and above pressure at rest; by contrast, exercise systemic vascular resistance added prognostic precision to vascular resistance at rest (P<.01). As for total mortality, systolic pressure and systemic vascular resistance at peak exercise carried prognostic information that was independent of the results at rest (P<.05); this was not the case for measurements during submaximal exercise. We conclude that the prognostic importance of blood pressure is related to systemic vascular resistance. The prognostic precision of exercise pressure, on top of pressure at rest, is limited. Exercise systemic vascular resistance, however, provides prognostic information beyond that available from measurements at rest, particularly for the incidence of cardiovascular events.
High blood pressure is a significant predictor of cardiovascular events and death.1 2 Because of the relatively weak relationship between conventionally measured blood pressure at rest and outcome, there has been a continuous search for improvement of the predictive power of blood pressure, eg, by use of ambulatory blood pressure monitoring3 4 5 or pressure measurements during exercise.6 7 8 9 However, the prognostic value of the hemodynamic components of blood pressure has not been explored except in one report in which outcome was associated with systemic vascular resistance assessed by use of noninvasive methods in the resting subject.10 In the present study, we sought to determine which of the hemodynamic variables, when measured both at rest and during exercise by use of reliable invasive methods, would be of paramount importance for the prediction of cardiovascular events and death in hypertensive patients. This is of particular interest for measurements made during exercise because an excessive elevation of blood pressure during exercise has been associated with a worse prognosis in population-based studies8 9 but not in selected hypertensive patients.7 It was hypothesized in one of the former studies9 that the prognostic value of an excessive blood pressure rise could result from a blunted decrease in systemic vascular resistance during exercise. In hypertensive individuals, however, a lesser exercise-induced reduction of systemic vascular resistance could be opposed by an attenuated increase of cardiac output and would therefore not result in an excessive rise of blood pressure. A different response of cardiac output could explain divergent findings in different study populations. We therefore conducted a long-term (median, 16.2 years) follow-up study in 143 male hypertensive patients in whom intra-arterial pressure and invasive cardiac output had been measured at rest and during bicycle exercise during the period 1972-1982. The present report is an extension of an earlier 11-year follow-up of these patients,7 in which only blood pressure was considered.
The study population consisted of 143 men who were referred for investigation of hypertension and who underwent invasive hemodynamic measurements at rest and during exercise between 1972 and 1982.7 Only patients with sinus rhythm and without evidence of ischemic or valvular heart disease, heart failure, claudication, renal insufficiency, or pulmonary disease were eligible.
Exercise Protocol and Hemodynamic Measurements
All hemodynamic measurements were performed in the morning, a few days after hospital admission, in one laboratory, where room temperature was 18° to 22°C as described earlier.11 Patients gave informed consent after the nature of the procedures had been explained. The brachial artery was cannulated for measurement of intra-arterial pressure (millimeters of mercury) and sampling of arterial blood. A venous catheter (Swan-Ganz, 93.110.5 Fr) was introduced into the antecubital vein and positioned in the pulmonary artery for sampling of mixed venous blood. Pressures were registered on a recorder (Siemens Mingograph 81). Uptake of oxygen (Vo2) and carbon dioxide output (Vco2) were measured continuously by the open-circuit method and expressed as liters per minute (standard temperature pressure dry). Minute-volume (body temperature pressure saturated) was determined by a pneumotachograph and oxygen and carbon dioxide concentrations by paramagnetic and infrared gas analyzers, respectively. The respiratory gas exchange ratio was calculated as Vco2/Vo2. Cardiac output (liters per minute) was determined by the direct oxygen Fick method and divided by body surface area (cardiac index; liters per minute per meter squared). Systemic vascular resistance index was calculated from mean brachial artery pressure, obtained by electrical damping, and cardiac index and expressed as units per meter squared. Heart rate (beats per minute) was recorded from the electrocardiogram.
A first set of hemodynamic measurements was obtained with patients in supine rest 30 minutes after the technical procedures. The patients were then seated on an electromagnetically braked bicycle ergometer, and rest sitting measurements were obtained 10 minutes later. A graded uninterrupted exercise test was subsequently started at a workload of 20 W for 4 minutes; the load was increased by 30 W every 4 minutes until exhaustion. Hemodynamic measurements were performed during the last minute of every other exercise step and at the final workload. Data at rest with patients in the supine and sitting positions at 50 W (the highest workload performed by all patients) and at the final workload are used for analysis.
After the baseline examination, the patients were referred to their usual source of care. Their vital status was determined in 1994 through contacts with municipal authorities. Causes of death were ascertained from contacts with physicians or family members and from hospital files and autopsy reports if available. The health status of living patients was determined through a standardized questionnaire filled in by physicians or, in the absence of a response, through a shorter questionnaire, which could be filled in by the patients; if necessary, patients were contacted by telephone. In addition, the charts of patients followed at the University Hospitals in Leuven were checked for possible cardiovascular events. When cardiovascular events had occurred, the responsible physicians were contacted and all available documents concerning the events were checked. Events were coded according to the Ninth Revision of the International Classification of Diseases. The following events were considered: sudden death, myocardial infarction, cerebrovascular accident, heart failure, angina pectoris, transient ischemic attack, and peripheral vascular disease. Objective evidence was required for acceptance. The vital status of all patients could be determined in 1994. However, two living patients could not be traced, so their data from the 1989 survey7 were used in the present analysis.
Database management and statistical analysis were performed with the SAS software (SAS Institute Inc). Data are reported as mean±SD or median and range. Two categories of end points were considered: (1) all-cause mortality and (2) fatal and nonfatal cardiovascular events combined. In patients with more than one cardiovascular event, only the first event was considered. The survival analysis was performed by Cox regression12 13 and involved four steps. In a first step, we related outcome to age and age squared. Then, we assessed the prognostic importance of the hemodynamic variables at rest and during exercise with adjustment for age and also age squared when the quadratic term was significant. Third, we studied the independence of the prognostic significance of the hemodynamic variables during exercise from data at rest by inclusion of the respective values at sitting rest before exercise in the various models. Finally, we assessed whether the prognostic importance of hemodynamic variables was independent of traditional risk factors, that is, serum total cholesterol, smoking habits, electrocardiographic voltage, and body mass index. Only significant predictors of outcome in separate analyses were entered in the multivariate models.
The age of the 143 men ranged from 16 to 64 years at the time of the hemodynamic study and averaged 35.3±12.0 years. Mean height and weight were 174±7 cm and 77.9±12.5 kg, respectively. Blood pressure on admission to the hospital averaged 173±28/109±20 mm Hg. Antihypertensive treatment had been interrupted for at least 2 weeks in 95 patients, whereas the others had never been treated. Most patients had essential hypertension. Five patients had a renal artery stenosis, which was considered to be significant in 3, and 10 had evidence of renal parenchymal disease but with normal renal function. Routine investigations revealed no or only mild to moderate target-organ damage except in 2 patients, who had experienced a cerebrovascular accident 2 and 7 years before the study but were not limited in their capacity to exercise. Electrocardiographic voltages averaged 0.54±0.39 mV for the R wave in lead aVL, 1.34±0.69 for SV1, 1.88±0.74 for RV5, and 3.76±1.25 for the sum of the three. Half of the patients (51%) did not smoke at the time of the hemodynamic study. Total serum cholesterol averaged 5.46±1.09 mmol/L and serum creatinine 100±16 μmol/L. Hemodynamic data at rest and during exercise are given in Table 1⇓. Brachial intra-arterial pressure at supine rest ranged from 104 to 240 mm Hg for systolic pressure and from 46 to 118 mm Hg for diastolic pressure. Peak workload averaged 158±44 W and peak oxygen uptake 2.28±0.60 L/min or 29.5±7.8 mL/min per kilogram; the respiratory gas exchange ratio was 1.06±0.11 at the highest workload.
Events During Follow-up
The follow-up time in individual patients, until death or until the date of the last available information on their vital status in 1994, ranged from 0.7 to 21.6 years (median, 16.2 years); the total follow-up time was 2186 patient years. Seventeen patients died between 0.7 and 20.1 years (median, 4.7) after the hemodynamic study, 13 from cardiovascular and 4 from noncardiovascular causes (Table 2⇓). Mean age at the time of death was 51 years (range, 37 to 78). Thirty-eight patients experienced at least one cardiovascular event, of which 9 were fatal and 29 nonfatal (Table 2⇓). The 1 patient with nonfatal heart failure had been admitted to the hospital with pulmonary edema. Nine of the 10 patients with angina pectoris had undergone coronary arteriography, and greater than or equal to 50% narrowing of the luminal diameter of one or more coronary arteries had been found in all; the remaining patient had shown ischemic electrocardiographic changes during an anginal attack and suffered a fatal myocardial infarction 7 months later. One transient ischemic attack had been diagnosed by a physician on the basis of a transient neurological deficit in one arm and the presence of carotid artery atheromatotic lesions. Peripheral arterial disease was confirmed by stenotic lesions on arteriography in 5 patients and nonpalpable foot arteries in 2. Mean age at the time of the first cardiovascular event was 49 years (range, 35 to 72); the events occurred between 0.2 and 17.7 years (median, 7.3) after the hemodynamic evaluation. Nine patients had a second cardiovascular event and another 2 suffered at least 3 events; additional events were not considered in the analysis unless fatal.
Predictors of Cardiovascular Events
Age at baseline was curvilinearly (P<.001) correlated with the incidence of cardiovascular events. The model, which required both the linear (P<.001) and quadratic (P<.001) terms of age, showed that age did not carry any excess risk below 40 years. However, the risk ratios for a difference in age at baseline from 40 to 49 and from 50 to 59 years were 1.08 (0.72 to 1.64) and 3.92 (1.43 to 10.7), respectively. The prognostic importance of the hemodynamic variables at supine rest is summarized in Table 3⇓. The relative hazards rates, adjusted for age and age squared, were significant for systolic pressure and systemic vascular resistance index (P<.01) but not for cardiac index (P=.27). In addition, cardiac index did not add prognostic precision to that of mean blood pressure (P=.44). Fig 1⇓ summarizes the results with patients in the sitting position at rest and during exercise. Blood pressure, measured at 50 W and at peak workload, predicted the incidence of cardiovascular events before (P<.01 for systolic and P<.05 for diastolic pressures) but not after adjustment for pressures in the sitting position at rest (.10<P<.77). The prognostic importance of exercise systemic vascular resistance index (P<.01) persisted after adjustment for vascular resistance at sitting rest (P<.01). Exercise cardiac index tended to be inversely related to outcome; after adjustment for age, age squared, and cardiac index at rest, the relative hazards rates were 0.68 (P=.06) at 50 W and 0.84 (P=.14) at peak exercise (Fig 1⇓).
The sum of the electrocardiographic voltages RaVL, SV1, and RV5 was significantly related to the incidence of cardiovascular events after adjustment for age and age squared (relative hazards rate, 1.028 [1.002 to 1.054]; P<.05); this was not the case for smoking habits, serum cholesterol, and body mass index (.41<P<.52). The prognostic precision of the hemodynamic variables was not substantially altered by additional adjustment for the electrocardiographic voltages in multivariate regression analysis. The results at rest are given in Table 3⇑. Exercise systemic vascular resistance index remained significantly related to outcome after adjustment for age, age squared, vascular resistance at rest, and electrocardiographic voltages (relative hazards rates, 1.123 [1.018 to 1.239] at 50 W [P<.05] and 1.115 [1.035 to 1.201] at peak exercise [P<.01]).
Predictors of Total Mortality
Total mortality was related to only the linear term of age (relative hazards rate, 1.061 [1.019 to 1.104]; P<.01). Age-adjusted relative hazards rates were significant for resting systolic (1.029 [1.009 to 1.049]; P<.01) and diastolic (1.037 [1.003 to 1.072]; P<.05) pressures and systemic vascular resistance index (1.046 [1.003 to 1.090]; P<.05) but not for cardiac index (P=.60); cardiac index at rest did not add prognostic precision to the mean blood pressure level (P=.85). Fig 2⇓ gives the results for hemodynamic data with patients at rest in the sitting position and at both 50 W and peak exercise. Systolic pressure during exercise predicted total mortality independently of age (P<.01). After additional adjustment for the pressure at sitting rest, only peak exercise pressure remained related to outcome (P<.05). The adjusted diastolic exercise pressures did not predict mortality (.18<P<.29). Also, cardiac index was not related to outcome (.86<P<.99). The age-adjusted relative hazards rate was significant for systemic vascular resistance index at both 50 W (P=.05) and peak exercise (P<.01). After additional adjustment for vascular resistance at sitting rest, peak exercise vascular resistance remained related to mortality (P<.05). Total mortality was not related to body mass index, smoking habits, serum cholesterol, or electrocardiographic voltage after adjustment for age.
The assessment of the prognostic value of an elevated blood pressure has traditionally been based on conventional, casual, or clinic blood pressure measurements at rest, without consideration of the hemodynamic components of blood pressure, ie, cardiac output and systemic vascular resistance. We assessed the outcome of 143 men who were investigated for hypertension during the period 1972-1982, when blood pressure was measured intra-arterially and cardiac output by the direct oxygen Fick technique both at rest and during graded bicycle exercise until exhaustion. During the 2186 patient years of follow-up, the incidence of cardiovascular events and all-cause mortality were significantly related to intra-arterial pressure and systemic vascular resistance.
At rest, cardiac output did not add independent prognostic precision to that of blood pressure. The prognostic importance of systemic vascular resistance is compatible with the fact that hypertension is in general characterized by a high systemic vascular resistance, whereas cardiac output is elevated in merely one third of young hypertensive individuals with mild blood pressure elevation.14 15 16 In addition, patients with a history of cardiac disease and possibly low cardiac output and impaired prognosis were excluded from the present study. The results confirm and strengthen those from Mensah et al,10 who estimated systemic vascular resistance index from noninvasive mean blood pressure and echocardiographic cardiac index in 193 men and women with uncomplicated essential hypertension. In their study, patients who suffered a clinical event during the 11.6-year follow-up had higher blood pressure and vascular resistance at baseline, but independence from other confounders, such as age, sex, and left ventricular mass was not assessed.10
Although admittedly blood pressure is related to morbidity and mortality, debate continues over which type of blood pressure is best related to the complications of hypertension.3 4 5 6 7 8 9 In a shorter follow-up of the present patient population, totaling 1573 patient years, we examined the prognostic significance of exercise blood pressure and concluded that blood pressure at 50 W, at 50% of peak exercise, and at peak workload did not add prognostic precision to the pressure at rest.7 These findings have been disputed recently,8 9 but there are important differences between these studies and our previous report7 : a large number of healthy middle-aged men versus a smaller number of referred hypertensive patients; noninvasive versus intra-arterial blood pressure measurements; a short versus a longer period of rest before exercise; a relatively steep exercise protocol versus progressive graded multistage exercise, as conventionally used for clinical purposes; and differences in the studied end points and statistical methods. The present analysis, based on continued follow-up of our hypertensive patients, supports our earlier conclusion7 that intra-arterial pressure at submaximal and peak bicycle exercise does not add prognostic precision to the pressure measured at rest before exercise, except for the small independent predictive value of peak systolic pressure for total mortality.
Because divergent findings between normotensive and hypertensive individuals could be due to different hemodynamic responses to exercise, we analyzed the hemodynamic components of blood pressure in the present report. As expected, cardiac output rose and systemic vascular resistance decreased from sitting rest to 50 W and to peak bicycle exercise.17 Systemic vascular resistance, when measured at both submaximal and peak exercise, predicted the incidence of cardiovascular events significantly. In contrast to blood pressure, their prognostic value was preserved after adjustment for vascular resistance at rest. These results indicate that an impaired reduction of systemic vascular resistance from rest to exercise carries independent prognostic information. The decrease of vascular resistance during dynamic exercise is caused by powerful metabolic arteriolar dilatation in the working muscles,18 which receive up to 80% of cardiac output during maximal bicycle exercise. A blunted decrease of systemic vascular resistance is most likely due to attenuated arteriolar dilatation in these vascular beds as a result of structural vascular abnormalities. It is reasonable to assume that vascular hypertrophy, a marker of more severe hypertensive disease, forms the link between the impaired reduction of vascular resistance during exercise and the higher incidence of cardiovascular events. The fact that the independent prognostic importance of a persistently high systemic vascular resistance during exercise is not expressed in the prognostic significance of exercise blood pressure can be explained by the tendency to an inverse association between the incidence of cardiovascular events and cardiac output, suggesting that an attenuated rise of cardiac output during exercise carries a worse prognosis. Exercise probably unmasks latent cardiac dysfunction in the patients with the worse prognosis, in which diastolic filling problems at higher heart rates and an impaired left ventricular functional reserve may be involved.19 20 21 A lesser reduction of systemic vascular resistance and a blunted rise of cardiac output would result in an apparently normal blood pressure response to exercise.
Our results may explain why exercise blood pressure seems to provide independent prognostic information in healthy middle-aged men8 9 and not in selected hypertensive patients. It is conceivable that the positive association between outcome and an excessive blood pressure elevation during exercise observed in the population-based samples resulted from an attenuated exercise-induced vasodilatation, as suggested previously.9 It can be argued that in contrast to hypertensive patients, healthy subjects have a normal cardiac output response to exercise. Consequently, an impaired reduction of systemic vascular resistance would not be opposed by a blunted rise of cardiac output and is therefore expressed in excessive blood pressure elevation.
The relationships between the hemodynamic variables and incidence of cardiovascular events were independent of electrocardiographic voltage. These findings are of particular interest because it has been suggested that the additional prognostic precision of exercise blood pressure could be mediated by a higher left ventricular mass,9 known to carry a worse prognosis in the general population22 23 and in hypertensive patients,24 regardless of whether mass was assessed by electrocardiography22 or echocardiography.23 24 The observation that the prognostic precision of systemic vascular resistance is independent of electrocardiographic voltage suggests that it is independent of left ventricular mass. This conclusion should be interpreted with caution, however, because left ventricular mass was only assessed by electrocardiography and not by a more sensitive method such as echocardiography.
Similar to the results on the incidence of cardiovascular events, all-cause mortality was significantly related to intra-arterial pressure and systemic vascular resistance and not to cardiac output. The results are less consistent for the measurements during exercise, when only systemic vascular resistance at peak effort carried prognostic information over and above that of vascular resistance at rest. Moreover, the prognostic importance of vascular resistance was not opposed by cardiac output, so the prognostic precision of peak exercise pressure was independent of pressure at rest. The lower number of deaths than of events and the inclusion of noncardiovascular causes should be taken into account when interpreting the findings on total mortality.
Judged from our results, the practical value of blood pressure measurements during routine exercise testing for more precise determination of the prognosis of hypertensive patients is likely to be limited. Only systolic intra-arterial pressure at peak exercise added some prognostic precision to the pressure at rest; noninvasive measurements are not reliable at peak effort.25 At the workload of 50 W, which should be achievable in most patients, and when conventional measurements of systolic pressure are reasonably accurate, blood pressure did not carry independent prognostic information. Cardiac output can be measured noninvasively and accurately during exercise, for example, by the carbon dioxide rebreathing technique.26 The present study suggests that it could be useful to include such measurements in the exercise testing of hypertensive patients, but it remains to be seen whether systemic vascular resistance based on noninvasive measurements would yield equally useful information.
This study was supported by the National Fund for Medical Research NFWO, Brussels, Belgium. R.H. Fagard is holder of the Prof A. Amery Chair in Hypertension Research, founded by Merck Sharp & Dohme (Belgium). The authors gratefully acknowledge the assistance of Yvette Toremans, Nicole Ausseloos, Jos Delsupehe, and Julien Romont.
- Received February 12, 1996.
- Accepted February 28, 1996.
Kreger RE, Kannel WB. Influence of hypertension on mortality. In: Amery A, Fagard R, Lijnen P, Staessen J, eds. Hypertensive Cardiovascular Disease: Pathophysiology and Treatment. The Hague, Netherlands: Martinus-Nijhoff Publishers; 1982:451-463.
Kannel WB, Stokes J III. Hypertension as a cardiovascular risk factor. In: Bulpitt CJ, ed. Handbook of Hypertension: Volume 6, Epidemiology of Hypertension. Amsterdam, Netherlands: Elsevier Science Publishers; 1985:15-34.
Perloff D, Sokolow M, Cowan RM, Juster RP. Prognostic value of ambulatory blood pressure measurements: further analysis. J Hypertens. 1989;7(suppl 3):S3-S10.
Verdecchia P, Porcellati C, Schillaci G, Borgioni C, Ciucci A, Battistelli M, Guerrieri M, Gatteschi C, Zampi I, Santucci A, Santucci C, Reboldi G. Ambulatory blood pressure: an independent predictor of prognosis in essential hypertension. Hypertension. 1994;24:793-801.
Fagard R. Exercise blood pressure and cardiovascular morbidity and mortality. Int J Sports Cardiol. 1994;3:9-12.
Fagard R, Staessen J, Thijs L, Amery A. Prognostic significance of exercise versus resting blood pressure in hypertensive men. Hypertension. 1991;17:574-578.
Filipovsky J, Ducimetière P, Safar ME. Prognostic significance of exercise blood pressure and heart rate in middle-aged men. Hypertension. 1992;20:333-339.
Mundal R, Kjeldsen SE, Sandvik L, Erikssen G, Thanlow E, Erikssen J. Exercise blood pressure predicts cardiovascular mortality in middle-aged men. Hypertension. 1994;24:56-62.
Mensah GA, Pappas TW, Koren MJ, Ulin RJ, Laragh JH, Devereux RB. Comparison of classification of the severity of hypertension by blood pressure level and by WHO criteria in the prediction of concurrent cardiac abnormalities and subsequent complications in essential hypertension. J Hypertens. 1993;11:1429-1440.
Fagard R, Bulpitt C, Lijnen P, Amery A. Response of the systemic and pulmonary circulation to converting-enzyme inhibition (captopril) at rest and during exercise in hypertensive patients. Circulation. 1982;65:33-39.
Cox DR. Regression models and life-tables. J R Stat Soc. 1972;B34:187-220.
Mattheus DE, Farewell V. Proportional Hazards Regression. Basel, Switzerland: S Karger; 1985:148-157.
Messerli FH, De Carvalho JGR, Christie B, Frohlich ED. Systemic and regional hemodynamics in low, normal and high cardiac output borderline hypertension. Circulation. 1978;58:441-448.
Fagard R, Staessen J, Amery A. Haemodynamic aspects of human essential hypertension. In: Zanchetti A, Mancia G, eds. Handbook of Hypertension: Pathophysiology of Hypertension. Amsterdam, Netherlands: Elsevier Science Publishers. In press.
Amery A, Julius S, Whitlock LS, Conway J. Influence of hypertension on the hemodynamic response to exercise. Circulation. 1967;36:231-237.
Shepherd JT, Vanhoutte PM. The Human Cardiovascular System: Facts and Concepts. New York, NY: Raven Press Publishers; 1980:91-106.
Tubau JF, Szlachnic J, Braun S, Massie BM. Impaired left ventricular functional reserve in hypertensive patients with left ventricular hypertrophy. Hypertension. 1989;14:1-8.
Kannel WB, Gordon T, Castelli WP, Margolis JR. Electrocardiographic left ventricular hypertrophy and risk of coronary heart disease: the Framingham Study. Ann Intern Med. 1970;72:813-822.
Koren MJ, Devereux RB, Casale PN, Savage DD, Laragh JH. Relation of left ventricular mass and geometry to morbidity and mortality in uncomplicated essential hypertension. Ann Intern Med. 1991;114:345-352.
Gould BA, Hornung RS, Altman DG, Cashman PM, Raftery EB. Indirect measurements of blood pressure during exercise testing can be misleading. Br Heart J. 1985;53:611-615.
Reybrouck T, Fagard R. Assessment of cardiac output at rest and during exercise by a carbon dioxide rebreathing method. In: Robertson JIS, Birkenhäger WH, eds. Cardiac Output Measurement. Beerse, Belgium: Janssen Biomedical Science Series; 1990:21-25.