Hemodynamic Effects of Vasodilators and Long-Term Response in Heart Failure

JACC Vol. 3, No.6 1521 June 1984:1521-30

REPORTS ON THERAPY

JOSEPH A. FRANCIOSA, MD, FACC, W. BRUCE DUNKMAN, MD, FACC,* CHERYL L. LEDDY, MD*

Philadelphia, Pennsylvania and Little Rock, Arkansas

Hemodynamic responses to vasodilators are commonly ml/min per kg (p < 0.01), and exercise duration also assessed when starting long-term vasodilator treatment increased by 1.8 ± 3.5 minutes (p < 0.01). Changes in in patients with chronic left ventricular failure, although maximal oxygen uptake, however, did not correlate with the relation between short- and long-term responses is short-term changes in pulmonary wedge pressure (cor• not established. Thus, short- and long-term hemody• relation coefficient [r] = - 0.14), cardiac index (r = namic responses to placebo and vasodilators (isosorbide - 0.01) or systemic vascular resistance (r = - 0.20). dinitrate, minoxidil and enalapril or captopril) were Long-term hemodynamic changes also failed to correlate measured and long-term clinical efficacy was assessed with changes in exercise capacity. Baseline hemody• by changes in exercise capacity after 1 to 5 months of namics, cardiac dimensions and left ventricular ejection vasodilator administration (plus digitalis and diuretic fraction before vasodilator administration all failed to agents) in 46 patients with New York Heart Association correlate with baseline exercise capacity or with long• fun~tional class II to IV heart failure caused by cardio• term changes in exercise capacity. myopathy. There were no significant changes in hemo• Thus, hemodynamic measurements at initiation or dynamics or exercise capacity during placebo treatment. during follow-up of vasodilator therapy do not relate to After initial doses and during long-term administration long-term clinical efficacy assessed by exercise capacity of vasodilator drugs, hemodynamics were significantly in patients with chronic left ventricular failure. There• improved. After long-term vasodilator treatment, max• fore, the rationale for making invasive hemodynamic imal oxygen uptake during exercise increased by 2.9 ± measurements before initiating long-term vasodilator 5.7 ml/min per kg from a control value of 14.1 ± 5.6 therapy for heart failure is questioned.

The widespread use of vasodilator agents in the management hemodynamic responses to initial doses of vasodilators are of congestive heart failure is based largely on the well doc• related to the long-term efficacy of these agents in patients umented hemodynamic efficacy of these agents (I ,2). Whereas with chronic congestive heart failure. Furthermore, hemo• hemodynamic measurements are of clear value in catego• dynamics and left ventricular function are usually assessed rizing the actions of specific drugs, there is no proof that at rest, but measurements at rest do not correlate with the functional capacity of patients with congestive heart failure (3-5). Whether pre-drug baseline measurements help pre• dict long-term responses to therapy is also unknown. Despite From the Cardiovascular Section, Veterans Administration Medical Center and The Hospital of the University of Pennsylvania, Philadelphia. these considerations, it has been suggested (6,7) that selec• Pennsylvania and the Cardiovascular Division. Veterans Administration tion and dose titration of vasodilators be based on acute Medical Center and University of Arkansas for Medical Sciences, Little hemodynamic responses to these drugs. Such an approach Rock, Arkansas. This study was supported in part by a grant from the Medical Research Service of the Veterans Administration. Washington. usually requires costly hospitalization of patients for the D.C. This paper was presented in part at the annual scientific sessions of performance of invasive hemodynamic studies to initiate the American College of Cardiology. New Orleans. Louisiana. March, long-term vasodilator therapy. The present study was de• 1983. Manuscript received August 29, 1983; revised manuscript received December 13, 1983, accepted December 21, 1983. signed to test the validity of this practice. We measured *Current address: Veterans Administration Medical Center, University hemodynamics before and after initial doses of vasodilator and Woodland Avenues, Philadelphia, Pennsylvania 19104. drugs, and examined the relation between these measure• Address for reprints: Joseph A. Franciosa, MD, Cardiovascular Di• vision, University of Arkansas for Medical Sciences, 4301 West Markham ments and the subsequent long-term effects of vasodilators Street, Mail Slot 532, Little Rock, Arkansas 72205. in patients with chronic left ventricular failure.

the carbon dioxide rebreathing method, an indirect Fick Methods procedure previously validated by us in patients with heart Study patients. Studies were performed in 47 men with failure (8,9). congestive heart failure caused by ischemic heart disease or Exercise protocol. Control measurements of pressures, primary myocardial disease. Ischemic heart disease was di• heart rate and cardiac output at rest were obtained at 30 agnosed from a history of documented acute myocardial minute intervals until two successive determinations varied infarction or a demonstration of obstructive coronary artery by less than 10%. On completion of baseline measurements disease by coronary arteriography. Primary myocardial dis• at rest, patients performed upright bicycle exercise to symp• ease was diagnosed when no other cause of heart failure tomatic maximum (dyspnea or fatigue without any other was apparent. Patients with primary pulmonary disease, limiting cause) using a Monark bicycle ergometer. During primary valvular heart disease or acute myocardial infarction exercise, hemodynamic variables were monitored and ex• within the preceding 3 months were excluded. Patients whose pired gas was collected for on-line measurement of oxygen exercise ability was limited by angina pectoris or any cause consumption, carbon dioxide production and ventilatory ex• other than fatigue or dyspnea were also excluded. Patients change ratio using techniques described in a previous study were required to have symptoms of exertional dyspnea or (3,10). Exercise was begun at a work load of 150 kilopond• fatigue attributable to cardiac disease for at least 3 months meters (kpm) and increased by 150 kpm every 4 minutes while being treated with digitalis and diuretic drugs. Ad• until exhaustion. The ventilatory exchange ratio is used to ditional selection criteria required objective evidence of left indicate onset of lactate production and achievement of an• ventricular dysfunction manifested as one of the following: aerobic threshold (II). By these criteria all patients achieved left ventricular internal end-diastolic dimension greater than anaerobic threshold on all exercise tests as indicated by a 2.7 cm/m2 by echocardiography; cardiothoracic ratio greater significant increase in ventilatory exchange ratio. than 50% by chest X-ray film and left ventricular ejection Initiation of vasodilator drug therapy. On completion fraction less than 46% by radionuclide angiography. Patients of all baseline measurements, patients were continued in meeting these initial screening criteria gave written informed the study if pulmonary wedge pressure was greater than 15 consent; the consent form and study protocol were approved mm Hg at rest or greater than 20 mm Hg during exercise, by the local Human Studies Committees. or if cardiac index at rest was not more than 2.5 liters/min 2 Hemodynamic study. Consenting patients were then per m . Patients meeting these criteria were then given their brought to a special procedure room adjacent to the medical first dose of vasodilator drug or placebo being tested. Iso• intensive care unit for baseline hemodynamic and exercise sorbide dinitrate was administered as a 40 mg oral dose to evaluation. Vasodilator therapy was discontinued at least eight patients, minoxidil as a 20 mg oral dose to eight 72 hours before hemodynamic studies. Diuretic drugs were patients, captopril as a 25 mg oral dose to eight patients withheld on the day of study, but maintenance digoxin doses and enalapril, a new angiotensin-converting enzyme inhib• were administered. Right heart catheterization was per• itor (12), was administered as a 5 or 10 mg oral dose to formed using a triple lumen Swan-Ganz thermistor-equipped eight patients. lsosorbide dinitrate and minoxidil were ad• catheter, which was inserted percutaneously by way of an ministered in fixed doses, and hemodynamic measurements antecubital or femoral vein and positioned in the pulmonary at rest were repeated 90 minutes after giving isosorbide artery. Pressures were measured with a Bell and Howell dinitrate and 4 hour after the dose of minoxidil; these are transducer positioned at the mid-chest level with the patient times of approximate peak effect based on our previous supine, and were displayed on a multichannel direct writing experience with these agents (13,14). Captopril and ena• Hewlett-Packard recorder. Right atrial and pulmonary artery lapril doses were titrated over 24 hours until pulmonary phasic, mean and occluded pressures were recorded. Pul• wedge pressure decreased or cardiac index increased by monary wedge pressure was taken as occluded pulmonary 30%, or to a maximal dose of 150 mg of captopril and 20 artery pressure or pulmonary artery diastolic pressure if this mg of enalapril. In three patients the dose of captopril was did not differ from the occluded pressure by more than 4 titrated to 25 mg, in four patients to 100 mg and in one mm Hg. These pressures were not interchanged in the same patient to 150 mg. Acute hemodynamic effects of captopril patient for data analysis. Systemic blood pressure was mea• were measured 2 hours after the last dose during titration. sured by the standard cuff technique, heart rate and rhythm In five patients the dose of enalapril was titrated to 10 mg were continuously recorded electrocardiographically and and in three patients to 20 mg. Acute hemodynamic effects cardiac output was measured by thermodilution in at least of enalapril were taken as those observed 4 hours after the triplicate using iced saline solution (4°C) with a carbon last dose during titration, based on our prior experience with dioxide-powered injector and an Edwards Laboratories car• this agent (15). Placebo, identical in appearance to isosor• diac output computer. In three patients, the Swan-Ganz bide dinitrate or minoxidil, was administered to IS patients catheter could not be maintained in satisfactory position in and hemodynamic values were measured I 1/2 to 4 hours the pulmonary artery, and cardiac output was measured by later. JACC Vol. 3. No.6 FRANCIOSA ET AL. 1523 June 1984:1521-30 VASODILATORS IN HEART FAILURE

Maintenance drug therapy. After completion of base• Mean pUlmonary artery and wedge pressures were elec• line evaluation, patients continued with the dose of vaso• tronically integrated, whereas mean systemic arterial pres• dilator drug tested. Isosorbide dinitrate was administered 40 sure was calculated as diastolic pressure plus one-third of mg four times daily; minoxidil as 20 mg twice daily; cap• the pulse pressure. topril as 25 mg three times daily in three patients, 100 mg Statistical analysis was performed using standard for• three times daily in four patients and 150 mg three times mulas for mean, standard deviation and linear regression. daily in one patient and enalapril was given as 10 mg twice Values before and after test drugs were compared by paired daily in five patients and 20 mg twice daily in three patients. t tests. Long-term response to therapy was assessed by changes Placebo was administered four times daily in eight patients in exercise capacity. Correlation coefficients (r) between and twice daily in seven. Study drugs and digoxin were long-term responses and baseline measurements were cal• maintained constant throughout the follow-up period while culated. Because of the large number of r values calculated, diuretic drugs were adjusted as needed. Patients were seen the significance level for these was adjusted for the number regularly at I to 4 week intervals. Patients taking placebo, of tests performed, thus requiring a probability (p) value of minoxidil and isosorbide dinitrate were followed up for 3 0.05 divided by the number of t tests (Bonferroni correc• months, those taking enalapril for 1 month and those taking tion). In this manner, a p value less than 0.001 was required captopril for 2 to 5 months. At the end of the follow-up for an r value to be significant. period, patients returned to the laboratory for repeat rest The assignment of ~pecific agents to individual patients and exercise measurements using these same techniques. was not by any prospective scheme but was largely arbitrary. Measurements were made at the same time after test drug Most of the patients were undergoing baseline evaluation dosing as during the baseline study. During follow-up, one before entering trials of long-term vasodilator therapy for patient taking minoxidil died from severe heart failure before heart failure. repeat studies could be completed. The 46 patients who completed the follow-up period are discussed in this report. Analysis of data. Cardiac index was calculated as car• Results diac output divided by body surface area and expressed as Clinical classifications. The vasodilator-treated group 2 liters/min per m . Systemic vascular resistance was calcu• consisted of 31 men whose age (± standard deviation) lated as mean arterial pressure minus right atrial pressure averaged 55.8 ± 8.1 years (range 40 to 72) (Table 1). The divided by cardiac output and expressed in arbitrary units. cause of congestive heart failure was ischemic heart disease

Table 1. Baseline Comparison of Patients With Congestive Heart Failure Treated With Vasodilators or Placebo

Vasodilators Placebo p Value

Patients (no.) 31 15 Age (yr) 55.5 ± 8.1 57.3 ± 7.0 NS Cause of heart failure (% of patients) Ischemic heart disease 55 60 NS Primary myocardial disease 45 40 NS Clinical class* (% of patients) II 29 20 NS III 55 67 NS IV 16 13 NS Cardiothoracic ratio (%) 55.2 ± 7.1 57.0 ± 4.1 NS Left ventricular diastolic 65.1 ± 10.5 67.1 ± 5.4 NS dimension (mm) Left ventricular ejection 28.7 ± 17.8 27.2 ± 15.0 NS fraction (%) Heart rate (beats/min) 77.5 ± 12.3 81.5 ± 13.9 NS Mean arterial pressure (mm Hg) 90.2 ± 13.7 87.6 ± 10.1 NS Pulmonary wedge pressure (mm Hg) 20.2 ± 9.2 22.6 ± 6.5 NS 2 Cardiac index (liters/min per m ) 2.36 ± 0.70 2.26 ± 0.52 NS Systemic vascular resistance 21.1 ± 7.8 21.1 ± 5.4 NS (units) Duration of follow-up (mo) 2.4 ± 1.1 2.9 ± 0.2 NS

Values shown are mean ± I standard deviation. * = according to criteria of the New York Heart Association. NS = not significant; p = probability. 1524 FRANCIOSA ET AL. JACC Vol. 3. No.6 VASODILATORS IN HEART FAILURE June 1984:1521-30

in 17 patients and primary myocardial disease in 14 patients. did not meet hemodynamic criteria at rest had marked in- All patients were symptomatic. New York Heart Associa- creases in pulmonary wedge pressure to greater than 20 mm tion functional class II symptoms were present in 9 patients, Hg during exercise. Of the three patients without baseline class III symptoms in 17 patients and class IV symptoms pulmonary wedge pressure measurements, two had a low in 5 patients. cardiac index. The other patient (Patient 11) had a normal Baseline cardiac dimensions and hemodynamic data of cardiac index at rest, but was included in the study because these patients are shown in Table 2. All but eight patients of marked exercise intolerance and a very low (16%) left had a cardiothoracic ratio greater than 50%. Of these eight ventricular ejection fraction at rest. The placebo-treated group, patients, seven had a left ventricular ejection fraction below which was included because of the possibility of sponta- 40% and the remaining patient met the screening criterion neous changes in hemodynamic status or exercise capacity of increased left ventricular diastolic dimension. The group during vasodilator administration, did not differ from the as a whole had increased cardiothoracic ratio and left ven- vasodilator-treated group in any baseline characteristic (Ta- tricular diastolic dimension with depressed left ventricular ble I). ejection fraction indicative of left ventricular dysfunction. Hemodynamic drug responses. Initial and later follow- Baseline hemodynamic measurements also indicated ab- up hemodynamic responses are summarized in Figure I. normal left ventricular function with elevated pulmonary Mean arterial pressure decreased by an average of 10.0 ± wedge pressure and low cardiac index. Four patients who 12.5 mm Hg (p < 0.00 I) after initial vasodilator doses and

Table 2. Baseline Characteristics of Patients With Chronic Left Ventricular Failure

Case Drug CTR LVOO LVEF HR MAP PWP CI SVR

1 IS 50 77 19 64 70 28 3.89 10 2 IS 44 56 39 80 105 9 1.87 31 3 IS 58 64 22 96 97 36 1.70 29 4 IS 63 80 103 17 1.86 26 5 IS 60 75 75 29 1.24 32 6 IS 57 74 24 97 87 30 1.54 29 7 IS 55 75 104 26 2.52 24 8 IS 52 84 95 26 1.83 22 9 M 51 52 71 65 119 9 3.81 13 10 M 49 56 35 76 104 6 3.01 20 II M 58 71 16 60 82 3.45 12 12 M 61 61 31 92 103 7 3.14 15 13 M 61 67 21 59 88 17 1.94 17 14 M 42 59 35 77 80 2.03 19 15 M 50 58 27 65 86 1.86 26 16 E 53 62 45 95 101 II 2.46 20 17 E 55 50 40 66 104 8 2.50 19 18 E 66 66 12 95 98 23 2.10 20 19 E 60 70 26 74 108 17 3.32 18 20 E 60 96 10 74 76 27 2.17 18 21 E 72 75 28 87 80 30 2.42 13 22 E 40 64 47 62 98 20 3.29 16 23 E 55 53 32 84 100 4 2.76 24 24 C 61 24 89 100 38 1.62 41 25 C 52 64 25 81 77 26 2.16 17 26 C 64 57 23 80 70 19 1.83 18 27 C 50 60 33 50 90 15 1.85 27 28 C 47 63 25 80 79 22 2.44 13 29 C 54 33 92 76 17 1.50 39 30 C 52 69 19 67 67 24 2.14 16 31 C 60 84 12 81 74 25 2.81 II Mean 55.2 65.1 28.7 77.5 90.2 20.2 2.36 21.1 ± SO 7.1 10.5 12.8 12.3 13.7 9.2 0.70 7.8 ± SEM 1.3 2.1 2.5 2.2 2.5 1.7 0.13 1.4

2 C = captopril; CI = cardiac index (liters/min per m ); CTR = cardiothoracic ratio (%); E = enalapril; HR = heart rate (beats/min); IS = isosorbide dinitrate; LVOD = left ventricular diastolic dimension (mm); LVEF = left ventricular ejection fraction (%); M = minoxidil; MAP = mean arterial pressure (mm Hg); PWP = mean pulmonary wedge pressure (mm Hg); SD = standard deviation; SEM = standard error of the mean; SVR = systemic vascular resistance (units). JACC Vol. 3, No.6 FRANClOSA ET AL. 1525 June 1984:1521-30 VASODILATORS IN HEART FAILURE

120 35 MAP PWP (mmHg) (mmHg) 100 25 L..·L--l Figure 1. Hemodynamic responses to vasodilators in patients with chronic left ventricular failure. C 80 15 = control; CI = cardiac index; I = initial response ~. after dose titration; L = late responses after I to 3 ~ months; MAP = mean arterial pressure; PWP = 60 5 pulmonary wedge pressure; SVR = systemic vas• cular resistance; * = p < 0.05; ** = P < 0.02; *** = P < 0.01; **** = P < 0.001 (all p values 4 35 are compared with original baseline control values). CI SVR Solid circles and lines = mean ± standard error (L/min/m2) (units) of the mean for vasodilator-treated group. Open 3 25 circles and broken lines = mean ± standard error of the mean for placebo-treated group. 2 ------r------r 15 ~ ......

5 C L C L

was still significantly decreased by 6.9 ± 11.0 mm Hg at cantly (- 10.9 ± 12.0 mm Hg, p < 0.05), and cardiac later follow-up (p < 0.01). Paralleling changes in blood index increased after enalapril administration, systemic vas• pressure, systemic vascular resistance decreased by 4.0 ± cular resistance was significantly reduced by 2.8 ± 2.4 5.9 units after initial vasodilator administration (p < 0.00 I), units (p < 0.02). Thus, each test drug individually produced and at a later follow-up study was still reduced by 4.5 ± significant hemodynamic changes of some type. 4.4 units (p < 0.001). Thus, the active test drugs did produce Vasodilator effect on exercise capacity. The effects of immediate vasodilation that was sustained during long-term long-term vasodilator administration on exercise capacity treatment. There were no significant hemodynamic changes are shown in Figure 2. Exercise duration increased by 1.8 observed any time after placebo administration. ± 3.5 minutes (p < 0.01) from a control value of 9.6 ± Changes in pulmonary wedge pressure and cardiac index 5.2 minutes. This was equivalent to a 47% increase (p < reflected the known actions of the individual drugs tested. 0.01). Maximal oxygen uptake at baseline study averaged For the entire group pulmonary wedge pressure decreased 14.1 ± 5.6 mllmin per kg and increased by 2.9 ± 5.7 initially by an average of 5.9 ± 7.4 mm Hg (p < 0.001), mllmin per kg (29%) during long-term vasodilator admin• and at follow-up study was insignificantly reduced by 3.3 istration (p < 0.01). The average control value for maximal ± 8.7 mm Hg from baseline control levels. Cardiac index oxygen uptake is markedly below normal for men of this increased by 0.28 ± 0.69 liters/min per m2 after initial age and places these patients in our exercise class III (3). vasodilator administration (p < 0.05), and remained sig• In the placebo-treated group, exercise capacity was similar nificantly increased by 0.48 ± 0.89 liters/min per m2 at at baseline and did not change during follow-up. follow-up (p < 0.02). Pulmonary wedge pressure decreased The relation between long-term responses, as indicated by 7.5 ± 3.0 mm Hg after initial doses of isosorbide din• by changes in maximal oxygen uptake, and initial hemo• itrate (p < 0.001), a predominant venodilator, and was dynamic responses to vasodilator drugs are shown in Figure unchanged by the arterial dilator minoxidil. Enalapril, an 3. Maximal oxygen uptake was chosen as the indicator for arterial and venous dilator, reduced pulmonary wedge pres• long-term response because of its greater objectivity, es• sure slightly by 3.0 ± 7.5 mm Hg (not significant), and pecially when achievement of anaerobic threshold is doc• captopril decreased it by 10.8 ± 8.7 mm Hg (p < 0.001). umented. In our patients, exercise duration and maximal Captopril-treated patients did not have pulmonary wedge oxygen uptake were closely correlated (r = 0.84, P < pressure measurements at later follow-up, probably ac• 0.001). There were no significant correlations between counting for the insignificant reduction of wedge pressure changes in pulmonary wedge pressure, cardiac index or in the whole group of patients at follow-up. Cardiac index systemic vascular resistance after initial vasodilator doses increased significantly only after captopril and minoxidil and long-term changes in maximal oxygen uptake. Because administration. Because blood pressure decreased signifi- II patients required a change in diuretic therapy during 1526 FRANCIOSA ET AL. JACC Vol. 3. No.6 VASODILATORS IN HEART FAILURE June 1984:1521-30

VASODILATORS PLACEBO VASODILATORS PLACEBO 28 Ip(OOI) In 5 ) 40 Ip(OOI) (n 5)

24 "",35 l:), ~ .!;: ""'.!;: 20 ~ 30 Figure 2. Effects of long-term vasodilator and ~ ~" <: / placebo treatment on exercise capacity in patients Cl l<.J with chronic left ventricular failure. Treated = i::: 16 :.c 25 ~ values after I to 5 months of treatment. Solid ~ Q.. circles and lines = mean ± standard error of the is ~ mean for the vasodilator-treated group. Open cir- 12 t} 20 ~ ~ c1es and lines = mean ± standard error of the ~ + ~ mean for the placebo-treated group. :..: ei Cl + ~ 8 -J 15 ~ + ~ 'q 4 ~ 10

0 5 Control Treated Control Treated Control Treated Control Treated follow-up and I of these also had a transient reduction in increased 1.04 ± 1.37 liters/min per m2 (p < 0.01) and minoxidil dosage, data were analyzed separately in the other systemic vascular resistance was reduced - 5.6 ± 5.4 units 20 patients who had no changes in any of their medications (p < 0.05). At follow-up in enalapril-treated patients, mean during the follow-up period. In these 20 patients, there was arterial pressure and systemic vascular resistance were re• still no significant correlation between changes in pulmonary duced (-9.3 ± 7.5 mm Hg [p < 0.01] and -2.0 ± 1.9 wedge pressure and in maximal oxygen uptake (r = 0.01), units [p < 0.02], respectively). Follow-up hemodynamic or between changes in cardiac index and in maximal oxygen measurements were not obtained in captopril-treated pa• uptake (r = 0.33). Likewise, initial changes in systemic tients. Maximal oxygen consumption increased during treat• vascular resistance also failed to correlate significantly with ment with each vasodilator, except minoxidil, but there were long-term changes in maximal oxygen uptake. Not only no significant correlations between any initial hemodynamic were there no significant correlations between initial hemo• effect and changes in exercise capacity in any individual dynamic effects of vasodilator drugs and long-term re• vasodilator-treated subgroup (Table 3). sponses assessed by exercise capacity, the long-term hemo• Baseline cardiac performance, baseline exercise ca• dynamic responses and changes in exercise capacity also pacity and changes in vasodilator therapy. Because re• failed to correlate. The r values for long-term hemodynamic sponsiveness to a therapeutic intervention could relate to responses and changes in maximal oxygen uptake were: the severity of heart failure at baseline, we performed linear - 0.38 for pulmonary wedge pressure, - 0.07 for cardiac regression analysis in an attempt to relate various baseline index and - 0.23 for systemic vascular resistance. These characteristics of the patient group to exercise capacity at poor correlations between hemodynamic responses and baseline and to changes in exercise capacity at later follow• changes in exercise capacity were not altered when data up study (Table 4). Using the statistical criteria set forth by were analyzed using percent change instead of absolute us and requiring a p value less than 0.001, none of the r changes for all variables, as was done in our preliminary values shown in the table achieved statistical significance. report (16). These observations showing lack of correlation between Comparison of individual vasodilator drugs. Although baseline hemodynamic characteristics and baseline exercise individual vasodilator-treated SUbgroups were small, they capacity confirm our previous findings (3,4). The observed were analyzed individually. As previously stated, each vaso• lack of correlation between these same baseline variables dilator produced some short-term hemodynamic effect. These and long-term response further suggests that baseline as• effects were sustained over time as a reduction in pulmonary sessment of severity of heart failure based on hemodynamic wedge pressure -7.8 ± 6.8 mm Hg (p < 0.02) and sys• criteria is not very predictive of long-term therapeutic re• temic vascular resistance - 6.0 ± 4.7 units (p < 0.01) at sponses judged by changes in exercise tolerance. It is also follow-up in isosorbide dinitrate-treated patients. At follow• noteworthy that the baseline exercise capacity was also not up, in the minoxidil-treated group, cardiac index was still related to long-term changes in exercise capacity. JACC Vol. 3. No.6 FRANCI0SA ET AL. 1527 June 1984:1521-30 VASODILATORS IN HEART FAILURE

20 treated with primary arterial dilating agents, such as hy• • 14 I r = - . 1 dralazine or minoxidiL which more effectively increase car• 0...... 10 • • diac output. Patients with symptoms attributable to both low ~ 0. o· • cardiac output and pulmonary vascular congestion should • .0. ••0. .~ 0 0. •• 0 do better with agents that dilate both arteries and veins, such " 1· 0 • ~ 0. as prazosin, hydralazine plus nitrate, or angiotensin-con• ...... verting enzyme inhibitors. In addition to using hemody• ~ -10 -16 -12 -8 -4 0 4 8 12 namic measurements to select a vasodilator, it has also been ~ Change in PWP (mm Hg) common practice to titrate a chosen agent to a dose that ~ 20 • produces a desired hemodynamic response (6,20-22). Q... I r = -.01 I ~ Although these practices have been recommended, actual • 0. • ~ 10 experience has shown no correlation between acute hemo• •• 0. dynamic effects and long-term clinical responses (23-25). ~ 0 t::II. i,p. i 0 # 0. 0 Earlier studies (23) using hydralazine titrated to achieve an ~ 0 • •0 0. 0 "optimal" short-term increase in cardiac output found that -..J • ~ -10 0 long-term outcomes were variable and unrelated to the he• -I I 3 4 ~ -3 -2 0 2 modynamically titrated doses of hydralazine. Another study Change in CIlL/min/m2) ~ (24) using hydralazine and nitrates in combination also found 20 a poor relation between long-term clinical response and the .!:: • I r= -.20 I magnitude of initial hemodynamic responses. The authors •• 0. ~ 10 • observed, however, that lack of any short-term response c::: tl o..~ 0 • was associated with a poor long-term clinical response. Thus, (; .0. • 0. • 0 0 °i· 0 it appears that some quantifiable response is desirable but 0 o~ • • • the magnitude of that response is not predictive of later -I O~_"""",,_--,~_,",-:-o_--,:::__J,-_-±-_~ outcome. Therefore, selection of any dose thought to be -16 -12 -8 -4 0 4 8 12 hemodynamically effective should be sufficient for insti• Change in SVR (units) • IS oM .E 6C tution of long-term therapy . Relation ofshort- and long-term hemodynamic effects Figure 3. Relation between acute hemodynamic effects of vaso• to clinical efficacy. The earlier studies just cited were in• dilators and long-tenn changes in exercise capacity in patients with tended primarily to assess long-term efficacy of vasodilators chronic left ventricular failure. C = captopril; CI = cardiac index; E = enalapril; IS = isosorbide dinitrate; M = minoxidil; PWP in heart failure, and not to examine the relation between = pulmonary wedge pressure; SVR = systemic vascular resis• baseline hemodynamic evaluation and long-term response. tance. No r values are statistically significant. Values of r shown Furthermore, they all used hydralazine, which has been are for all groups combined; no r values within individual groups shown to be no more effective than placebo in altering are significant. exercise capacity during long-term administration to patients with chronic left ventricular failure (26). Our study repre• sents an initial attempt to relate baseline hemodynamic sta• Discussion tus, as well as short- and long-term hemodynamic effects Selection of optimal vasodilator agent. Hemodynamic of vasodilators, to long-term clinical efficacy in patients with measurements are commonly employed in the diagnosis and chronic left ventricular failure. Furthermore. we used a va• management of patients with congestive heart failure. Such riety of vasodilator drugs with primarily venous, primarily measurements have been valuable in characterizing and arterial or balanced arterial and venous dilating ability. Be• quantitating the effects of pharmacologic interventions, es• cause long-term clinical response is difficult to assess, we pecially the site, duration and magnitude of action of vas• chose the objective measurement of exercise capacity as an odilator drugs (2,17). Because hemodynamic abnormalities indicator of symptomatic status and clinical efficacy. Such are usually present in patients with congestive heart failure, testing has commonly been used to demonstrate clinical it has been assumed that pharmacologic correction of these efficacy of vasodilators in patients with heart failure abnormalities relates to long-term therapeutic responses. (21,22,27). We were unable to show any significant cor• Therefore, it has been advocated that selection ofvasodilator relations between baseline hemodynamic values. initial agents be tailored to suit the individual hemodynamic profile hemodynamic responses or late hemodynamic responses to (6,7,18,19). Patients with symptoms primarily of pulmo• the various vasodilators tested and long-term effects on ex• nary vascular congestion might benefit more from drugs with ercise capacity. Our results were probably not related to predominant preload-reducing effects, such as nitrates. In ineffective doses, as each of the drugs tested produced some contrast, symptoms of low cardiac output might be better acute beneficial hemodynamic response. Furthermore. vas- 1528 FRANClOSA ET AL. JACC Vol. 3. No.6 VASODILATORS IN HEART FAILURE June 1984:1521-30

Table 3. Relation Between Acute Hemodynamic Effects of Individual Vasodilators and Long• Term Effects on Exercise Capacity in Patients With Heart Failure

Isosorbide Dinitrate Minoxidil Enalapril Captopril Change in maximal oxygen 6.9 ± 2.4t -0.9 ± 4.3 1.6 ± 3.5* 3.3 ± 4.7* uptake (mlimin per kg) Coefficient of correlation between change in maximal oxygen uptake and initial change in: Pulmonary wedge pressure 0.45 0.33 -0.04 -0.12 Cardiac index 0.57 -0.36 -0.04 -0.47 Systemic vascular resistance -0.42 0.54 0.33 -0.27

Values shown = mean ± I standard deviation. * = p < 0.1; t = P < 0.05. No coefficients of correlation are statistically significant (all p > 0.05). odilator responses in the group were sustained as systemic reported (21,26-28). Effects on exercise capacity were lim• vascular resistance was still significantly reduced at late ited by the failure of minoxidil to increase maximal oxygen follow-up. Changes in pulmonary wedge pressure and car• uptake, despite producing sustained vasodilation. This re• diac output varied according to the specific vasodilator drug sponse to minoxidil is similar to that observed during admin• employed, as expected. istration of hydralazine, a similar arterial dilator, in heart Validity of hemodynamic effects and changes in ex• failure (26). ercise capacity. The overall effects of vasodilators on Despite these considerations, we did not observe signif• hemodynamics and exercise capacity in our patients might icant correlations between hemodynamic effects and changes be considered modest, but they are consistent with previous in exercise capacity within any individual vasodilator-treated reports (28) and probably reflect varying responses to in• subgroup. Therefore, we believe that our overall results are dividual agents. Thus, the lack of effect of minoxidil and valid. Furthermore, the changes in hemodynamics and ex• the modest effect of enalapril on pulmonary wedge pressure ercise capacity that occurred during vasodilator treatment combined to limit the overall effect on wedge pressure. It are thought to be significant not only on a statistical basis, is known that minoxidil has little or no effect on pulmonary but also because these variables remained unchanged during artery pressures, and the observed responses to enalapril are placebo administration to similar patients. consistent with those reported by others (14,29). Likewise, Correlation between baseline patient characterization the overall increase in cardiac output was limited by the and changes in exercise capacity. The present results con• modest effects of isosorbide dinitrate and enalapril, again firm our previous observation (34) that baseline indicators consistent with previous reports (2,15,29). In this regard, of severity of heart failure, including cardiac dimensions, although captopril consistently increases cardiac output, the hemodynamics and left ventricular ejection fraction, do not magnitude of this change has been relatively modest in correlate with objectively measured exercise capacity at that patients with heart failure (2,22,28). The observed changes time. These observations have now been extended to show in exercise duration and maximal oxygen uptake for the a further lack of correlation between these same baseline group as a whole were also within the range previously patient characteristics and changes in exercise capacity dur-

Table 4. Relation of Baseline Cardiac Performance to Baseline Exercise Capacity and Changes in Exercise Capacity During Vasodilator Treatment in Patients With Chronic Heart Failure

Coefficient of Correlation Coefficient of Correlation With Change in Exercise Baseline Characteristic With Baseline Exercise Capacity Capacity

Cardiothoracic ratio -0.47 0.34 Left ventricular diastolic dimension -0.26 0.41 Left ventricular ejection fraction 0.42 -0.45 Pulmonary wedge pressure -0.41 0.26 Cardiac index 0.51 -0.48 Systemic vascular resistance 0.38 0.42 Maximal oxygen uptake -0.36 No r values achieved statistical significance (p < 0.001). JACC Vol. 3, No.6 FRANCIOSA ET AL. 1529 June 1984:1521-30 VASODILATORS IN HEART FAILURE

ing long-term treatment. In addition, the symptomatic status in the setting of clinical investigation. However, the results of patients at baseline study, assessed by exercise testing, of our study suggest that it is not necessary in clinical prac• is unrelated to long-term changes in exercise capacity during tice to routinely perform expensive invasive hemodynamic treatment. Thus, baseline assessment of severity of symp• studies to initiate vasodilator therapy in patients with toms or of degree of cardiac dysfunction does not appear congestive heart failure. Physicians should select those agents to relate to long-term changes in functional capacity of pa• that are not contraindicated and with which they are most tients with chronic left ventricular failure during therapeutic familiar. trials of vasodilator drugs. These results should be applied only to patients with chronic heart failure as ample clinical experience suggests that immediate hemodynamic improve• ment is frequently associated with relief of symptoms in References patients with acute left ventricular failure (30-34). In fact, I. Cohn IN. Franciosa lA. During therapy: vasodilator therapy of cardiac patients with more severe baseline hemodynamic abnor• failure. N Engl 1 Med 1977;297:27-31.254-8. malities appear to respond better to vasodilator administra• 2. Franciosa lA. Cohn IN. Hemodynamic responsiveness to short- and long-acting vasodilators in left ventricular failure. Am 1 Med tion in the setting of acute heart failure (2,33,34). 1978:65: 126-33. Clinical implications. The present results are clinically 3. Franciosa lA. Functional capacity of patients with chronic left ven• relevant as they raise questions concerning the need to hos• tricular failure. Relationship ofbicycle exercise performance to clinical pitalize patients for costly and prolonged invasive hemo• and hemodynamic characterization. Am 1 Med 1979:67:461-6. dynamic studies to initiate long-term vasodilator therapy for 4. Franciosa lA. Park M, Levine TB. Lack of correlation between ex• ercise capacity and indices of resting left ventricular performance in chronic congestive heart failure. It is no longer our practice heart failure. Am 1 Cardiol 1981;47:33-9. to routinely evaluate hemodynamic status at baseline. It is 5. Benge lW. Litchfield RL. Marcus ML. Exercise capacity in patients essential, however, that patients be started on a presumed with severe left ventricular dysfunction. Circulation 1980:61 :955-9. hemodyna.mically effective dose of one or more vasodila• 6. Packer M. Meller 1. Medina N. Godin R. Herman MV. Dose re• tors. In the vast majority of patients, the effective dose range quirements of hydralazine in patients with severe chronic congestive heart failure. Am 1 Cardiol 1980;45:655-60. for most vasodilators is well known (2,17,22,27), and it is 7. Rubin SA, Chatterjee K, Parmley WW. Metabolic assessment of ex• only necessary that patients be started on these doses without ercise in chronic heart failure patients treated with short-term vaso• quantitating hemodynamic responses. Hemodynamic mea• dilators. Circulation 1980;61 :543-8. surements may still be needed in some situations. They may 8. Franciosa lA, Ragan DO, Rubenstone S1. Validation of the CO2 be required to confirm the diagnosis of heart failure because rebreathing method for measuring cardiac output in patients with hy• pertension or heart failure. 1 Lab Clin Med 1976;88:672-82. the symptoms of dyspnea and fatigue are very nonspecific 9. Franciosa lA. Evaluation of the CO2 rebreathing cardiac output method and noninvasive measurements of cardiac size and function in seriously ill patients. Circulation 1977:55:449-55. may be inconclusive. Measurement of left ventricular filling 10. Franciosa lA. Cohn IN. Effect of isosorbide dinitrate on response to pressure and cardiac output at rest and during exercise may submaximal and maximal exercise in patients with congestive heart definitively establish the diagnosis of heart failure, Hemo• failure. Am 1 Cardiol 1979;43:1009-14. dynamic measurements may also be useful to look for evi• II. Naimark A, Wasserman K, Mcilroy MB. Continuous measurement of ventilatory exchange ratio during exercise. 1 Appl Physiol dence of resistance or tolerance to drugs when therapeutic 1964: 19:644-52. response has been unsatisfactory (35,36), 12. Ferguson RK. Vlasses PH. Swanson BL, et al. Effects of enalapril, We do not wish to suggest that hemodynamic measure• a new converting enzyme inhibitor, in hypertension. Clin Pharmacol ments have no value in assessing long-term outcome in Ther 1982;32:48-53. patients with chronic congestive heart failure. This study 13. Franciosa lA. Mikulic E. Cohn IN, lose E. Fabie A. Hemodynamic effects of orally administered isosorbide dinitrate in patients with shows only that such measurements do not relate to long• congestive heart failure. Circulation 1974:50: 1020-4. term outcome as assessed by exercise capacity, which pre• 14. Franciosa lA. Cohn IN. Effects of minoxidil on hemodynamics in sumably reflects symptomatic status of patients. Most im• patients with congestive heart failure. Circulation 1982;63:652-7. portantly, there is no evidence to suggest that measured 15. Dunkman WB, Wilen M, Franciosa lA. Enalapril (MK-421). a new exercise capacity relates to long-term prognosis, that is, angiotensin converting enzyme inhibitor: acute and chronic effects in heart failure. Chest 1983;84:539-45. mortality and morbidity complicating the course of heart 16. Franciosa lA, Wilen M. Relationship between acute hemodynamic failure. It may not, therefore, be satisfactory to rely only responses to vasodilators and long-term efficacy in heart failure (abstr). on results of exercise testing as a means for establishing 1 Am Coli Cardiol 1983;1:713. efficacy of treatment. Survival of patients with congestive 17. Franciosa lA, Pierpont G. Cardiovascular clinical pharmacology of heart failure is related to measurements of left ventricular impedance reducing agents. 1 Chronic Dis 1981 ;34:341-52. function and hemodynamics, survival being reduced as these 18. Mason DT. Afterload reduction and cardiac performance. Physiologic basis of systemic vasodilators as a new approach in treatment of indexes worsen (37,38). Therefore, such measurements are congestive heart failure. Am 1 Med 1978:65: 106-25. important in evaluating overall efficacy of therapeutic in• 19. Packer M. Is there a drug of choice for vasodilator therapy in patients terventions in heart failure and can justifiably be included with refractory heart failure? (editorial). Chest 1980;77:7-9. 1530 FRANCIOSA ET AL. JACC Vol. 3, NO.6 VASODILATORS IN HEART FAILURE June 1984:1521-30

20. Chatterjee K, Ports TA, Brundage BH, Massie B, Holly AN, Parmley and chronic hemodynamic evaluations. J Am Coli Cardiol WW. Oral hydralazine in chronic heart failure: sustained beneficial 1983;2:865-71. hemodynamic effects. Ann Intern Med 1980;92:600-4. 30. Hamosh P, Cohn IN. Left ventricular function in acute myocardial 21. Colucci WS, Wynne J, Holman BL, Braunwald E. Long-term therapy infarction. J Clin Invest 1971 ;50:523-33. of heart failure with prazosin: a randomized double blind trial. Am J 31. Forrester JS, Diamond G, Chatterjee K, Swan HJC. Medical therapy Cardiol 1980;45:337-44. of acute myocardial infarction by application ofhemodynamic subsets. 22. Levine TB, Franciosa JA, Cohn IN. Acute and long-term response to N Engl J Med 1976;295:1356-62, 1404-13. an oral converting-enzyme inhibitor, captopril, in congestive heart 32. Guiha NH, Cohn IN, Mikulic E, Franciosa JA. Limas CJ. Treatment failure. Circulation 1980;62:35-41. of refractory heart failure with infusion of nitroprusside. N Engl Med 23. Packer M, Meller J, Medina N, Gorlin R, Herman MY. Importance 1974;291 :587-92. of left ventricular chamber size in determining the response to hy- 33. Franciosa JA, Guiha NH, Limas CJ, Rodriguera E, Cohn IN. Im• dralazine in severe chronic heart failure. N Engl J Med 1980;303:250-5. proved left ventricular function during nitroprusside infusion in acute 24. Massie B, Ports T, Chatterjee K, et al. Long-term vasodilator therapy myocardial infarction. Lancet 1972; 1:650-7. for heart failure: clinical response and its relationship to hemodynamic 34. Chatterjee K, Parmley WW, Ganz W, et al. Hemodynamic and met• measurements. Circulation 1981 ;63:269-78. abolic responses to vasodilator therapy in acute myocardial infarction. 25. Walsh WF, Greenberg BH. Results of long-term vasodilator therapy Circulation 1973;48: 1183-93. in patients with refractory heart failure. Circulation 1981;64:499-505. 35. Thadani U, Manyari D, Parker JO, Fung H. Tolerance to the circu• 26. Franciosa JA, Weber KT, Levine TB, et al. Hydralazine in the long- latory effects of oral isosorbide dinitrate. Rate of development and term treatment ofchronic heart failure: lack ofdifference from placebo. cross-tolerance to glyceryl dinitrate. Circulation 1980;61 :526-35. Am Heart J 1982;104:587-94. 36. Packer M, Meller J, Gorlin R, Herman MY. Hemodynamic and clin• 27. Franciosa JA, Goldsmith SA, Cohn IN. Contrasting acute and chronic ical tachyphylaxis to prazosin-mediated afterload reduction in severe effects of nitrates on exercise capacity in heart failure. Am J Med chronic congestive heart failure. Circulation 1979;59:531-9. 1980;69:559-66. 37. Fuster Y, Gersh BJ, Giuliani ER, Tajik AJ, Brandenburg RO, Frye 28. Franciosa JA. Effectiveness of long-term vasodilator administration RL. The natural history of idiopathic dilated cardiomyopathy. Am J in the treatment of chronic left ventricular failure. Prog Cardiovasc Cardiol 1981 ;47:525-3I. Dis 1982;24:319-30. 38. Franciosa JA, Wilen M, Ziesche S, Cohn IN. Survival in men with 29. DiCarlo L, Chatterjee K, Parmley WW, et al. Enalapril: a new an• chronic left ventricular failure due to either coronary heart disease or giotensin-converting enzyme inhibitor in chronic heart failure: acute idiopathic dilated cardiomyopathy. Am J Cardiol 1983;51:831-6.