- Email: [email protected]
Use of aspirin in conjunction with angiotensin-converting enzyme inhibitors does not worsen long-term survival in heart failure
International Journal of Cardiology 88 (2003) 207–214 www.elsevier.com / locate / ijcard
Use of aspirin in conjunction with angiotensin-converting enzyme inhibitors does not worsen long-term survival in heart failure Kishore J. Harjai*, Sergio Solis, Ananth Prasad, Jill Loupe Department of Cardiology, Ochsner Clinic, New Orleans, USA Received 15 November 2001; received in revised form 20 June 2002; accepted 16 July 2002
Abstract Background: A negative interaction has been shown to exist between aspirin and angiotensin-converting enzyme inhibitors (ACE-I) in subjects with heart failure. We explored the effect of combined ACE-I and aspirin therapy compared to ACE-I without aspirin on clinical outcomes in patients with heart failure. Methods: 430 consecutive subjects (70614 years, 55% male, 41% with coronary artery disease) released from the hospital with a primary diagnosis of heart failure were classified into three groups based on the use of aspirin and ACE-I at discharge: ACE-I without aspirin (group I, n5134), ACE-I with aspirin (group II, n5138) and no ACE-I (group III, n5158). Follow-up (all-cause mortality and the composite end-point of mortality or emergent heart transplant) was available in 406 (94%) patients at a median duration of 28 months. Differences in outcomes between patient groups were compared using contingency tables, Kaplan–Meier survival, and Cox regression analyses. Similar analyses were conducted in four predefined subsets (patients with and without coronary artery disease, and those with left ventricular ejection fraction #45%, and .45%). Results: Death and the composite end-point occurred in 155 (38%) and 165 (41%) patients, respectively. In the total cohort as well as in the four subsets, the treatment group showed no association with clinical outcomes in univariate or multivariate analyses. Conclusions: In patients with a principal discharge diagnosis of heart failure, the use of aspirin, in combination with ACE-I, does not worsen long-term survival compared to the use of ACE-I without aspirin. 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Heart failure; Aspirin; Angiotensin converting enzyme inhibitor
1. Introduction Symptomatic heart failure is a progressive condition characterized by a poor long-term prognosis. Annual mortality in excess of 50% is reported in subjects with severe heart failure [1]. Several studies have shown an impressive reduction in medium to long-term mortality with the use of angiotensin-converting enzyme inhibitors (ACE-I) [2]. *Corresponding author. William Beaumont Hospital, 3601 W. 13 Mile Road, Royal Oak, MI 48304, USA. Tel.: 11-248-551-5000; fax: 11-248551-4299. E-mail address: [email protected] (K.J. Harjai).
Aspirin, by virtue of its antiplatelet effects, is known to be beneficial in the primary and secondary prevention of coronary artery disease [3–5]. Thus, it would seem prudent to combine aspirin with ACE-I in the management of patients with heart failure. However, a negative interaction has been shown to exist between ACE-I and aspirin in heart failure. Hall et al. showed an attenuation of the beneficial hemodynamic effects of ACE-I (lower systemic and pulmonary vascular resistance, lower left ventricular filling pressure and higher cardiac output) by prior or concomitant administration of aspirin [6]. Aspirin has also been shown to negate the beneficial ventilatory
0167-5273 / 02 / $ – see front matter 2002 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0167-5273(02)00401-1
208
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
and endothelial effects of ACE-I as well [7]. Further, aspirin attenuates the effect of antihypertensive medications [8–10]. By interfering with glomerular filtration and sodium and water excretion, aspirin is also known to blunt the effect of diuretic agents [11]. These adverse effects of aspirin are suspected to be mediated by its antiprostaglandin actions. We performed this study in subjects discharged alive from the hospital with a principal discharge diagnosis of heart failure. We explored the effect of combined ACE-I and aspirin therapy compared to ACE-I without aspirin on clinical outcomes in patients with heart failure.
2. Methods
2.1. Inclusion and exclusion criteria Over a consecutive 12-month period from September 1, 1994 to August 31, 1995, a total of 614 hospitalizations with a principal discharge diagnosis of heart failure (diagnosis related group 127) occurred at the Ochsner Foundation Hospital in 447 patients. The first hospitalization in each patient was considered the index hospitalization for the purpose of this analysis. We excluded patients who were less than 21 years of age (n56) and those who did not survive the index hospitalization (n511). Thus, 430 patients constituted the study group.
2.2. Data collection Demographic and clinical data at the time of hospitalization was obtained from a database maintained by the Utilization Review Department at the hospital. This database tracks patient age, gender, race, length of stay, need for intensive care stay, and specialty of the caregiver physician (classified for this study as cardiologist or noncardiologist), among several other variables. Medical records were reviewed to obtain initial blood pressures and heart rate readings at the time of hospitalization. Admission serum chemistries (serum sodium, serum creatinine, blood urea nitrogen) and discharge medications were extracted from databases maintained by the clinical laboratory and the in-hospital pharmacy, respectively.
Medications dispensed to patients on the day of discharge were designated discharge medications. Two-dimensional echocardiography had been performed in 306 (71%) patients either during or within 6 months preceding the index hospitalization. These studies were interpreted by experienced echocardiographers unrelated to the study design. Left ventricular ejection fraction was estimated visually (‘eyeball’ method); we and others have previously validated the accuracy of visual estimation of ejection fraction [12,13].
2.3. Patient classification Based on the use of ACE-I and aspirin at the time of discharge from the hospital, patients were classified into three different treatment groups: group I (ACE-I without aspirin, n5134), group II (ACE-I with aspirin, n5138), and group III (no ACE-I, n5158).
2.4. Study end-points The primary end-point of the study was all-cause mortality. The secondary end-point was the composite of death from any cause or emergent heart transplantation. Emergent heart transplantation was defined as transplant in a UNOS (United Network of Organ Sharing) status I patient. Study outcomes were determined from a combination of hospital and outpatient medical records and, if required, telephone interviews with the patients or their next of kin. Complete clinical follow-up was available in 406 (94.4%) patients at a median duration of 28 months (range 0.1–59 months). The proportion of patients lost to follow-up was similar between the three patient groups (6.0, 2.9 and 7.6%, respectively; P value not significant).
2.5. Statistics Continuous data are expressed as mean6one standard deviation, and discrete variables as percentages. A P value ,0.05 was considered to denote statistical significance. Differences between treatment groups were assessed using x 2 (discrete variables), or means tests (continuous variables), restricting analyses to two groups at a time. The incidence of study end-
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
points between treatment groups was assessed using x 2 test and Kaplan–Meier survival analyses. Differences between patients who met the study end-points versus those who did not were assessed using x 2 (discrete variables), or means tests (continuous variables). Cox proportional hazards regression was used to assess the presence of an independent relation between the treatment group (e.g. group I vs. II, or I vs. III, or II vs. III) and the study end-points. Clinical and demographic variables were selected for potential inclusion in multivariate analyses if they had a significant univariate association with the study outcome. Patient age, gender, race, and coronary artery disease were included as potential confounders in all multivariate analyses. Adjusted odds ratios and 95% confidence intervals were calculated for each of the covariates in multivariate analyses.
2.6. Study subgroups The study population was divided into patients with coronary artery disease (n5161), and those without coronary artery disease (n5245). Based on echocardiographic left ventricular systolic function (data available in 306 patients), two more subgroups of patients were defined: those with left ventricular ejection fraction 545% (n5186), and those with left ventricular ejection fraction .45% (n5120). Within each subgroup, an association was sought between treatment group and study outcomes, using the statistical methods described above. In analyses of subgroups of patients with and without left ventricular systolic dysfunction, echocardiographic variables were also tested for potential inclusion in the regression models.
3. Results
3.1. Baseline clinical characteristics Patients were a mean age of 70614 years; 237 (55%) patients were male and 289 (67%) were white. Coronary artery disease, hypertension, atrial fibrillation, and chronic obstructive pulmonary disease were present in 174 (41%), 243 (57%), 138 (32%), and 55 (13%) of patients, respectively. Table 1 shows a
209
comparison of the clinical characteristics between the three treatment groups. Compared to group I, group II patients were older, and had a higher incidence of coronary artery disease and hypertension. Group III patients were older than groups I and II, and had a lower incidence of left ventricular systolic dysfunction, higher serum creatinine and blood urea nitrogen values and were less likely to be under the primary care of a cardiologist during the index hospitalization.
3.2. Univariate correlates of study end-points The incidence of death and the composite endpoint of death or emergent heart transplant, and the median duration of follow-up in the entire cohort as well as in the subgroups of patients are shown in Table 2. Compared to survivors, patients who died during clinical follow-up were older, more likely to be white, have coronary artery disease, be discharged on nitrate therapy, and had a higher number of readmissions in the 6 months preceding the index hospitalization. Patients who died during follow-up also had higher serum creatinine and blood urea nitrogen, but lower serum sodium and diastolic blood pressure at the time of hospitalization. They had lower interventricular septum and posterior wall thickness on two-dimensional echocardiography. Compared to patients who survived without emergent heart transplant, those who died or had heart transplantation were more likely to be white, have coronary artery disease, low ejection fraction (545%); be discharged on nitrate therapy; and had higher number of readmissions in the 6 months preceding the index hospitalization. They had higher serum creatinine and blood urea nitrogen, but lower serum sodium, systolic blood pressure, and diastolic blood pressure at the time of hospitalization. They had higher left ventricular end-diastolic diameter, and lower interventricular septum and posterior wall thickness on two-dimensional echocardiography.
3.3. Comparison of patient groups 3.3.1. Total cohort Crude comparison of the incidence of study endpoints between treatment groups is shown in Fig. 1. Differences between patient groups were not statisti-
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
210
Table 1 Baseline characteristics of study patients Clinical characteristic
Demographics Age (years)† Male gender White Co-morbid conditions COPD Coronary artery disease ¶,§ Hypertension ¶ Atrial fibrillation
Group I (n5134) ACE-I without aspirin
Group II (n5138) ACE-I with aspirin
Group III (n5158) No ACE-I
67614 78 (58) 86 (64)
70614 80 (58) 93 (67)
71613 79 (50) 110 (70)
14 39 64 47
20 76 86 38
21 59 93 53
(11) (29) (48) (35)
(15) (55) (62) (28)
(13) (37) (59) (34)
Index hospitalization data Admission characteristics Systolic BP (mmHg)† Diastolic BP (mmHg) Heart rate (bpm) Serum sodium (mequiv. / l) Serum creatinine (mg / dl)†,§ Blood urea nitrogen (mg / dl)†,§ Length of stay Need for intensive care stay In-hospital care by cardiologist †,§
133628 78616 90622 13765 25616 1.560.7 4.163.0 11 (8.2) 89 (66)
140631 80619 88620 13764 24613 1.561.0 3.862.2 14 (10.2) 95 (69)
145634 80618 92623 13765 31621 2.161.6 4.263.7 21 (13.3) 87 (55)
Echocardiographic characteristics (n5306) Echocardiographic data available Ejection fraction545% §,† Left ventricular end-diastolic diameter (cm)§,† Left atrial diameter (cm)§,† Interventricular septum (cm) Left ventricular posterior wall (cm)§
86 (65) 63 / 86 (73) 5.661.1 5.060.8 1.0260.25 1.0560.23
104 (75) 73 / 104 (70) 5.461.0 4.960.8 1.0260.23 1.0160.23
116 (73) 50 / 116 (43) 5.061.1 4.660.9 1.0760.22 1.0860.21
COPD, chronic obstructive pulmonary disease; BP, blood pressure; numbers in parentheses indicate percentages. ¶, P,0.05 for Group I vs. Group II; §, P,0.05 for Group II vs. Group III. †, P,0.05 for Group I vs. Group III.
cally significant. Kaplan–Meier survival curves for the outcomes are shown in Fig. 2. No significant differences were seen between treatment groups for
either of the study end-points. In Cox regression analyses, the treatment group did not show any independent relation with either death or the compo-
Table 2 Incidence of study-end-points among patients discharged from the hospital with a principal diagnosis of heart failure Patients a
Death
Death or emergent heart transplant
Follow-up (months)b
All patients (n5406) Patients with CAD (n5161) Patients without CAD (n5245) Patients with ejection fraction #45% (n5176) Patients with ejection fraction .45% (n5115)
155 (38.2%) 76 (47.2%) 79 (32.3%) 73 (41.5%)
165 (40.6%) 82 (50.9%) 83 (33.9%) 79 (44.9%)
27.5 (0.1–59.0) 28 (1.0–59.0) 27 (0.1–57.0) 24 (0.5–59)
40 (34.8%)
40 (34.8%)
35 (0.1–57)
a b
Numbers in parentheses indicate the number of patients in each subgroup for whom follow-up data were available; CAD, coronary artery disease. Follow-up is expressed as median (range).
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
211
Fig. 1. Incidence of study end-points among patient groups. (a) All cause mortality (b) all cause mortality or emergent heart transplantation.
site end-point of death or emergent heart transplant. Clinical variables that showed a multivariate association with the study end-points are shown in Table 3.
3.3.2. Subgroup analyses Fig. 1 shows the crude incidence of the study end-points within each study subgroup. Treatment
212
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
group did not show a significant association with the study end-points in any of the subgroups in contingency tables, Kaplan–Meier survival, or Cox regression analyses. A weak trend towards lower mortality in group II versus group I was observed in the subgroup of patients with coronary artery disease (39.7 vs. 55.6%; P50.17). After multivariate adjustment, this trend remained insignificant (adjusted odds ratio for group II versus group I, 0.62; 95% CI 0.33–1.14; P50.13).
4. Discussion
Fig. 2. Kaplan–Meier survival estimates of study end-points between patient groups. (a) All cause mortality; (b) all cause mortality or emergent heart transplantation.
Our study results argue against a detrimental effect of aspirin on survival in symptomatic heart failure patients treated with angiotensin-converting enzyme inhibitors. Compared to patients who were discharged from the hospital on treatment with ACE-I without aspirin, those who were discharged on treatment with ACE-I and aspirin did not have worse survival. Similarly, aspirin use did not adversely affect the incidence of the composite end-point of death or emergent heart transplant either. On the contrary, in the subgroup of patients with CAD, the crude incidence of all-cause mortality (as well as all-cause mortality or emergent heart transplant) was nonsignificantly lower in patients treated with ACE-I combined with aspirin (Fig. 1), suggesting the possibility of a protective effect from this combination. The lack of detrimental effect from ACE-I and aspirin therapy compared to ACE-I without aspirin is
Table 3 Multivariate association of clinical variables with study outcomes Adjusted odds ratio a
95% confidence intervals
Death Number of hospitalizations in the prior 6 months Age (years) Diastolic blood pressure (mmHg) Group II vs. group I
1.29 1.03 0.986 1.18
1.13–1.47 1.02–1.05 0.974–0.998 0.73–1.92
Death or emergent heart transplantation Number of hospitalizations in the prior 6 months Age (years) Diastolic blood pressure (mmHg) Group II vs. group I
1.27 1.02 0.982 1.08
1.11–1.46 1.00–1.04 0.967–0.997 0.67–1.74
a
For continuous variables, odds ratios represent the change in the incidence of the outcome for each unit increase in the value of the variable.
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
consistent with recent posthoc analyses of the SOLVD and CONSENSUS II studies. In 6797 patients enrolled in the SOLVD prevention and treatment arms, among patients randomized to enalapril, the use of antiplatelet agents did not affect overall survival in this study [14]. Thus, in the enalapril arm, the adjusted hazards ratio for all-cause mortality among antiplatelet agent users versus nonusers was 1.00 (95% confidence intervals 0.85, 1.17). In a posthoc analysis of the Cooperative New Scandinavian Enalapril Survival Study II (CONSENSUS II), the interaction of enalapril with aspirin use was tested among 6090 patients within 24 h following an acute myocardial infarction. Among patients randomized to receive enalapril, those who also used aspirin at the time of enrollment into the study had lower 6-month mortality than those who did not [15]. Our findings and those of the SOLVD and CONSENSUS II investigators are suggestive that in patients with heart failure, there is no detrimental effect on mortality from the use of aspirin among users of ACE-I. In the same cohort of patients presented in this analysis, we have previously documented an increased risk of readmission for heart failure within 30 days after discharge among subjects discharged home on ACE-I in combination with aspirin, compared to those discharged on ACE-I without aspirin [16]. The lack of an adverse mortality effect from combined ACE-I and aspirin use might seem contradictory to increased early readmissions from the same combination. However, factors that affect heart failure readmission rates might not necessarily be operative in causing increased mortality. Further, it is also possible that the negative interaction between ACE-I and aspirin is most obvious in the early period after hospitalization, when patients are most susceptible to adverse hemodynamic interactions, and that it wanes with time.
4.1. Implications Experimental data are suggestive of the negative effects of aspirin coadministration on hemodynamic, ventilatory, and endothelial benefits provided by angiotensin-converting enzyme inhibition in heart failure [7,8]. On the other hand, observational clinical evidence, based on our study and posthoc analyses of SOLVD and CONSENSUS II studies, indicates that combined therapy with ACE-I and aspirin is not any
213
worse than ACE-I therapy without aspirin. A ‘negative interaction’ between antiplatelet agents and enalapril is nevertheless apparent in the SOLVD and CONSENSUS II studies, since randomization to enalapril (versus placebo) led to a survival benefit in nonusers of antiplatelet agents, but not so in users of antiplatelet agents in both these studies. However, in the current era of unquestioned benefit from ACE-I in heart failure [2], the clinically relevant question is whether addition of aspirin, in patients already receiving ACE-I, is beneficial, neutral, or detrimental to long-term mortality. Our results indicate that among patients on chronic ACE-I therapy, the addition of aspirin does not have any adverse effect on long-term mortality. Further, among patients with CAD, a trend towards improvement in mortality is seen in patients treated with the combination of ACE-I and aspirin, compared to those treated with ACE-I without aspirin. This finding is suggestive that combined aspirin and ACE-I use should be strongly considered in patients with symptomatic heart failure and CAD.
4.2. Limitations Our study suffers from the limitations inherent to nonrandomized, retrospective analyses. The dose or duration of therapy with either ACE-I or aspirin, or other agents known to be beneficial in heart failure, are not known. The study period preceded the publication of large-scale clinical studies on the beneficial effects of b-blocker therapy in symptomatic heart failure. As such, b-blocker therapy was generally avoided in patients with symptomatic heart failure; this may limit the validity of our results in the current era of more generous b-blocker use. Clinical follow-up was available in only 94.4% of the cohort. Furthermore, the number of patients in our study is relatively small, particularly in analyses of subgroups. Nevertheless, our results indicate that combined therapy with ACE-I and aspirin is not worse than ACE-I without aspirin.
Acknowledgements We are indebted to Mario Vaz for his assistance in the preparation of this manuscript.
214
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
References [1] Applefeld MM. Chronic congestive heart failure. Where have we been? Where are we heading. Am J Med 1986;80(Suppl 2B):73–7. [2] Garg R, Yusuf S. Overview of randomized trials of angiotensinconverting enzyme inhibitors on mortality and morbidity in patients with heart failure. J Am Med Assoc 1995;273:1450–6. [3] Second International Study of Infarct Survival (ISIS-2) Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both or neither among 17 187 cases of suspected acute myocardial infarction. Lancet 1988;2((8607)):349–60. [4] Hennekens CH, Buring JE, Sandercock P, Collins R, Peto R. Aspirin and other antiplatelet agents in the secondary and primary prevention of cardiovascular disease. Circulation 1989;80:749–56. [5] The Steering Committee of the Physicians’ Health Study Research Group. Final report on the aspirin component of the ongoing Physicians’ Health Study. New Engl J Med 1989;321:129–35. [6] Hall D, Zeitler H, Rudolph W. Counteraction of the vasodilator effects of enalapril by aspirin in severe heart failure. J Am Coll Cardiol 1992;20(7):1549–55. [7] Guazzi M, Marenzi G, Alimento M, Contini M, Agostoni P. Improvement of alveolar–capillary membrane diffusing capacity with enalapril in chronic heart failure and counteracting effect of aspirin. Circulation 1997;95:1930–6. [8] Oates JA. Antagonism of antihypertensive drug therapy by nonsteroidal anti-inflammatory drugs. Hypertension 1988;11:114–6, Review. [9] Houston MC. Nonsteroidal anti-inflammatory drugs and antihypertensives. Am J Med 1991;90:42S–7S.
[10] Johnson AG, Nguyen TV, Day RO. Do nonsteroidal anti-inflammatory drugs affect blood pressure? A meta-analysis. Ann Intern Med 1994;121:289–300. [11] Riegger GA, Kahles HW, Elsner D, Kromer EP, Kochsiek K. Effects of acetylsalicylic acid on renal function in patients with chronic heart failure. Am J Med 1991;90:571–5. [12] Harjai KJ, Nunez E, Turgut T, Shah MP, Humphrey JS, Newman J, Cheirif J, Smart FW, Ventura HO. The independent effects of left ventricular ejection fraction on short-term outcomes and resource utilization following hospitalization for heart failure. Clin Cardiol 1999;22:184–90. [13] Amico AF, Lichtenberg GS, Reisner SA, Stone CK, Schwartz RG, Meltzer RS. Superiority of visual versus computerized echocardiographic estimation of radionuclide left ventricular ejection fraction. Am Heart J 1989;118:1259–65. [14] Al-Khadra AS, Salem DN, Rand WM, Udelson JE, Smith JJ, Konstam MA. Antiplatelet agents and survival: a cohort analysis from the studies of left ventricular dysfunction (SOLVD) trial. J Am Coll Cardiol 1998;31:419–25. [15] Nguyen KN, Aursnes I, Kjekshus J. Interaction between enalapril and aspirin on mortality after acute myocardial infarction: subgroup analysis of the Cooperative New Scandinavian Enalapril Survival Study II. Am J Cardiol 1997;79:115–9. [16] Harjai KJ, Nunez E, Turgut T, Newman T. Effect of combined aspirin and angiotensin-converting enzyme inhibitor therapy versus angiotensin-converting enzyme inhibitor therapy alone on readmission rates in heart failure. Am J Cardiol 2001 15;87(4):483–7, A7.
Use of aspirin in conjunction with angiotensin-converting enzyme inhibitors does not worsen long-term survival in heart failure Kishore J. Harjai*, Sergio Solis, Ananth Prasad, Jill Loupe Department of Cardiology, Ochsner Clinic, New Orleans, USA Received 15 November 2001; received in revised form 20 June 2002; accepted 16 July 2002
Abstract Background: A negative interaction has been shown to exist between aspirin and angiotensin-converting enzyme inhibitors (ACE-I) in subjects with heart failure. We explored the effect of combined ACE-I and aspirin therapy compared to ACE-I without aspirin on clinical outcomes in patients with heart failure. Methods: 430 consecutive subjects (70614 years, 55% male, 41% with coronary artery disease) released from the hospital with a primary diagnosis of heart failure were classified into three groups based on the use of aspirin and ACE-I at discharge: ACE-I without aspirin (group I, n5134), ACE-I with aspirin (group II, n5138) and no ACE-I (group III, n5158). Follow-up (all-cause mortality and the composite end-point of mortality or emergent heart transplant) was available in 406 (94%) patients at a median duration of 28 months. Differences in outcomes between patient groups were compared using contingency tables, Kaplan–Meier survival, and Cox regression analyses. Similar analyses were conducted in four predefined subsets (patients with and without coronary artery disease, and those with left ventricular ejection fraction #45%, and .45%). Results: Death and the composite end-point occurred in 155 (38%) and 165 (41%) patients, respectively. In the total cohort as well as in the four subsets, the treatment group showed no association with clinical outcomes in univariate or multivariate analyses. Conclusions: In patients with a principal discharge diagnosis of heart failure, the use of aspirin, in combination with ACE-I, does not worsen long-term survival compared to the use of ACE-I without aspirin. 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Heart failure; Aspirin; Angiotensin converting enzyme inhibitor
1. Introduction Symptomatic heart failure is a progressive condition characterized by a poor long-term prognosis. Annual mortality in excess of 50% is reported in subjects with severe heart failure [1]. Several studies have shown an impressive reduction in medium to long-term mortality with the use of angiotensin-converting enzyme inhibitors (ACE-I) [2]. *Corresponding author. William Beaumont Hospital, 3601 W. 13 Mile Road, Royal Oak, MI 48304, USA. Tel.: 11-248-551-5000; fax: 11-248551-4299. E-mail address: [email protected] (K.J. Harjai).
Aspirin, by virtue of its antiplatelet effects, is known to be beneficial in the primary and secondary prevention of coronary artery disease [3–5]. Thus, it would seem prudent to combine aspirin with ACE-I in the management of patients with heart failure. However, a negative interaction has been shown to exist between ACE-I and aspirin in heart failure. Hall et al. showed an attenuation of the beneficial hemodynamic effects of ACE-I (lower systemic and pulmonary vascular resistance, lower left ventricular filling pressure and higher cardiac output) by prior or concomitant administration of aspirin [6]. Aspirin has also been shown to negate the beneficial ventilatory
0167-5273 / 02 / $ – see front matter 2002 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0167-5273(02)00401-1
208
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
and endothelial effects of ACE-I as well [7]. Further, aspirin attenuates the effect of antihypertensive medications [8–10]. By interfering with glomerular filtration and sodium and water excretion, aspirin is also known to blunt the effect of diuretic agents [11]. These adverse effects of aspirin are suspected to be mediated by its antiprostaglandin actions. We performed this study in subjects discharged alive from the hospital with a principal discharge diagnosis of heart failure. We explored the effect of combined ACE-I and aspirin therapy compared to ACE-I without aspirin on clinical outcomes in patients with heart failure.
2. Methods
2.1. Inclusion and exclusion criteria Over a consecutive 12-month period from September 1, 1994 to August 31, 1995, a total of 614 hospitalizations with a principal discharge diagnosis of heart failure (diagnosis related group 127) occurred at the Ochsner Foundation Hospital in 447 patients. The first hospitalization in each patient was considered the index hospitalization for the purpose of this analysis. We excluded patients who were less than 21 years of age (n56) and those who did not survive the index hospitalization (n511). Thus, 430 patients constituted the study group.
2.2. Data collection Demographic and clinical data at the time of hospitalization was obtained from a database maintained by the Utilization Review Department at the hospital. This database tracks patient age, gender, race, length of stay, need for intensive care stay, and specialty of the caregiver physician (classified for this study as cardiologist or noncardiologist), among several other variables. Medical records were reviewed to obtain initial blood pressures and heart rate readings at the time of hospitalization. Admission serum chemistries (serum sodium, serum creatinine, blood urea nitrogen) and discharge medications were extracted from databases maintained by the clinical laboratory and the in-hospital pharmacy, respectively.
Medications dispensed to patients on the day of discharge were designated discharge medications. Two-dimensional echocardiography had been performed in 306 (71%) patients either during or within 6 months preceding the index hospitalization. These studies were interpreted by experienced echocardiographers unrelated to the study design. Left ventricular ejection fraction was estimated visually (‘eyeball’ method); we and others have previously validated the accuracy of visual estimation of ejection fraction [12,13].
2.3. Patient classification Based on the use of ACE-I and aspirin at the time of discharge from the hospital, patients were classified into three different treatment groups: group I (ACE-I without aspirin, n5134), group II (ACE-I with aspirin, n5138), and group III (no ACE-I, n5158).
2.4. Study end-points The primary end-point of the study was all-cause mortality. The secondary end-point was the composite of death from any cause or emergent heart transplantation. Emergent heart transplantation was defined as transplant in a UNOS (United Network of Organ Sharing) status I patient. Study outcomes were determined from a combination of hospital and outpatient medical records and, if required, telephone interviews with the patients or their next of kin. Complete clinical follow-up was available in 406 (94.4%) patients at a median duration of 28 months (range 0.1–59 months). The proportion of patients lost to follow-up was similar between the three patient groups (6.0, 2.9 and 7.6%, respectively; P value not significant).
2.5. Statistics Continuous data are expressed as mean6one standard deviation, and discrete variables as percentages. A P value ,0.05 was considered to denote statistical significance. Differences between treatment groups were assessed using x 2 (discrete variables), or means tests (continuous variables), restricting analyses to two groups at a time. The incidence of study end-
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
points between treatment groups was assessed using x 2 test and Kaplan–Meier survival analyses. Differences between patients who met the study end-points versus those who did not were assessed using x 2 (discrete variables), or means tests (continuous variables). Cox proportional hazards regression was used to assess the presence of an independent relation between the treatment group (e.g. group I vs. II, or I vs. III, or II vs. III) and the study end-points. Clinical and demographic variables were selected for potential inclusion in multivariate analyses if they had a significant univariate association with the study outcome. Patient age, gender, race, and coronary artery disease were included as potential confounders in all multivariate analyses. Adjusted odds ratios and 95% confidence intervals were calculated for each of the covariates in multivariate analyses.
2.6. Study subgroups The study population was divided into patients with coronary artery disease (n5161), and those without coronary artery disease (n5245). Based on echocardiographic left ventricular systolic function (data available in 306 patients), two more subgroups of patients were defined: those with left ventricular ejection fraction 545% (n5186), and those with left ventricular ejection fraction .45% (n5120). Within each subgroup, an association was sought between treatment group and study outcomes, using the statistical methods described above. In analyses of subgroups of patients with and without left ventricular systolic dysfunction, echocardiographic variables were also tested for potential inclusion in the regression models.
3. Results
3.1. Baseline clinical characteristics Patients were a mean age of 70614 years; 237 (55%) patients were male and 289 (67%) were white. Coronary artery disease, hypertension, atrial fibrillation, and chronic obstructive pulmonary disease were present in 174 (41%), 243 (57%), 138 (32%), and 55 (13%) of patients, respectively. Table 1 shows a
209
comparison of the clinical characteristics between the three treatment groups. Compared to group I, group II patients were older, and had a higher incidence of coronary artery disease and hypertension. Group III patients were older than groups I and II, and had a lower incidence of left ventricular systolic dysfunction, higher serum creatinine and blood urea nitrogen values and were less likely to be under the primary care of a cardiologist during the index hospitalization.
3.2. Univariate correlates of study end-points The incidence of death and the composite endpoint of death or emergent heart transplant, and the median duration of follow-up in the entire cohort as well as in the subgroups of patients are shown in Table 2. Compared to survivors, patients who died during clinical follow-up were older, more likely to be white, have coronary artery disease, be discharged on nitrate therapy, and had a higher number of readmissions in the 6 months preceding the index hospitalization. Patients who died during follow-up also had higher serum creatinine and blood urea nitrogen, but lower serum sodium and diastolic blood pressure at the time of hospitalization. They had lower interventricular septum and posterior wall thickness on two-dimensional echocardiography. Compared to patients who survived without emergent heart transplant, those who died or had heart transplantation were more likely to be white, have coronary artery disease, low ejection fraction (545%); be discharged on nitrate therapy; and had higher number of readmissions in the 6 months preceding the index hospitalization. They had higher serum creatinine and blood urea nitrogen, but lower serum sodium, systolic blood pressure, and diastolic blood pressure at the time of hospitalization. They had higher left ventricular end-diastolic diameter, and lower interventricular septum and posterior wall thickness on two-dimensional echocardiography.
3.3. Comparison of patient groups 3.3.1. Total cohort Crude comparison of the incidence of study endpoints between treatment groups is shown in Fig. 1. Differences between patient groups were not statisti-
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
210
Table 1 Baseline characteristics of study patients Clinical characteristic
Demographics Age (years)† Male gender White Co-morbid conditions COPD Coronary artery disease ¶,§ Hypertension ¶ Atrial fibrillation
Group I (n5134) ACE-I without aspirin
Group II (n5138) ACE-I with aspirin
Group III (n5158) No ACE-I
67614 78 (58) 86 (64)
70614 80 (58) 93 (67)
71613 79 (50) 110 (70)
14 39 64 47
20 76 86 38
21 59 93 53
(11) (29) (48) (35)
(15) (55) (62) (28)
(13) (37) (59) (34)
Index hospitalization data Admission characteristics Systolic BP (mmHg)† Diastolic BP (mmHg) Heart rate (bpm) Serum sodium (mequiv. / l) Serum creatinine (mg / dl)†,§ Blood urea nitrogen (mg / dl)†,§ Length of stay Need for intensive care stay In-hospital care by cardiologist †,§
133628 78616 90622 13765 25616 1.560.7 4.163.0 11 (8.2) 89 (66)
140631 80619 88620 13764 24613 1.561.0 3.862.2 14 (10.2) 95 (69)
145634 80618 92623 13765 31621 2.161.6 4.263.7 21 (13.3) 87 (55)
Echocardiographic characteristics (n5306) Echocardiographic data available Ejection fraction545% §,† Left ventricular end-diastolic diameter (cm)§,† Left atrial diameter (cm)§,† Interventricular septum (cm) Left ventricular posterior wall (cm)§
86 (65) 63 / 86 (73) 5.661.1 5.060.8 1.0260.25 1.0560.23
104 (75) 73 / 104 (70) 5.461.0 4.960.8 1.0260.23 1.0160.23
116 (73) 50 / 116 (43) 5.061.1 4.660.9 1.0760.22 1.0860.21
COPD, chronic obstructive pulmonary disease; BP, blood pressure; numbers in parentheses indicate percentages. ¶, P,0.05 for Group I vs. Group II; §, P,0.05 for Group II vs. Group III. †, P,0.05 for Group I vs. Group III.
cally significant. Kaplan–Meier survival curves for the outcomes are shown in Fig. 2. No significant differences were seen between treatment groups for
either of the study end-points. In Cox regression analyses, the treatment group did not show any independent relation with either death or the compo-
Table 2 Incidence of study-end-points among patients discharged from the hospital with a principal diagnosis of heart failure Patients a
Death
Death or emergent heart transplant
Follow-up (months)b
All patients (n5406) Patients with CAD (n5161) Patients without CAD (n5245) Patients with ejection fraction #45% (n5176) Patients with ejection fraction .45% (n5115)
155 (38.2%) 76 (47.2%) 79 (32.3%) 73 (41.5%)
165 (40.6%) 82 (50.9%) 83 (33.9%) 79 (44.9%)
27.5 (0.1–59.0) 28 (1.0–59.0) 27 (0.1–57.0) 24 (0.5–59)
40 (34.8%)
40 (34.8%)
35 (0.1–57)
a b
Numbers in parentheses indicate the number of patients in each subgroup for whom follow-up data were available; CAD, coronary artery disease. Follow-up is expressed as median (range).
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
211
Fig. 1. Incidence of study end-points among patient groups. (a) All cause mortality (b) all cause mortality or emergent heart transplantation.
site end-point of death or emergent heart transplant. Clinical variables that showed a multivariate association with the study end-points are shown in Table 3.
3.3.2. Subgroup analyses Fig. 1 shows the crude incidence of the study end-points within each study subgroup. Treatment
212
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
group did not show a significant association with the study end-points in any of the subgroups in contingency tables, Kaplan–Meier survival, or Cox regression analyses. A weak trend towards lower mortality in group II versus group I was observed in the subgroup of patients with coronary artery disease (39.7 vs. 55.6%; P50.17). After multivariate adjustment, this trend remained insignificant (adjusted odds ratio for group II versus group I, 0.62; 95% CI 0.33–1.14; P50.13).
4. Discussion
Fig. 2. Kaplan–Meier survival estimates of study end-points between patient groups. (a) All cause mortality; (b) all cause mortality or emergent heart transplantation.
Our study results argue against a detrimental effect of aspirin on survival in symptomatic heart failure patients treated with angiotensin-converting enzyme inhibitors. Compared to patients who were discharged from the hospital on treatment with ACE-I without aspirin, those who were discharged on treatment with ACE-I and aspirin did not have worse survival. Similarly, aspirin use did not adversely affect the incidence of the composite end-point of death or emergent heart transplant either. On the contrary, in the subgroup of patients with CAD, the crude incidence of all-cause mortality (as well as all-cause mortality or emergent heart transplant) was nonsignificantly lower in patients treated with ACE-I combined with aspirin (Fig. 1), suggesting the possibility of a protective effect from this combination. The lack of detrimental effect from ACE-I and aspirin therapy compared to ACE-I without aspirin is
Table 3 Multivariate association of clinical variables with study outcomes Adjusted odds ratio a
95% confidence intervals
Death Number of hospitalizations in the prior 6 months Age (years) Diastolic blood pressure (mmHg) Group II vs. group I
1.29 1.03 0.986 1.18
1.13–1.47 1.02–1.05 0.974–0.998 0.73–1.92
Death or emergent heart transplantation Number of hospitalizations in the prior 6 months Age (years) Diastolic blood pressure (mmHg) Group II vs. group I
1.27 1.02 0.982 1.08
1.11–1.46 1.00–1.04 0.967–0.997 0.67–1.74
a
For continuous variables, odds ratios represent the change in the incidence of the outcome for each unit increase in the value of the variable.
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
consistent with recent posthoc analyses of the SOLVD and CONSENSUS II studies. In 6797 patients enrolled in the SOLVD prevention and treatment arms, among patients randomized to enalapril, the use of antiplatelet agents did not affect overall survival in this study [14]. Thus, in the enalapril arm, the adjusted hazards ratio for all-cause mortality among antiplatelet agent users versus nonusers was 1.00 (95% confidence intervals 0.85, 1.17). In a posthoc analysis of the Cooperative New Scandinavian Enalapril Survival Study II (CONSENSUS II), the interaction of enalapril with aspirin use was tested among 6090 patients within 24 h following an acute myocardial infarction. Among patients randomized to receive enalapril, those who also used aspirin at the time of enrollment into the study had lower 6-month mortality than those who did not [15]. Our findings and those of the SOLVD and CONSENSUS II investigators are suggestive that in patients with heart failure, there is no detrimental effect on mortality from the use of aspirin among users of ACE-I. In the same cohort of patients presented in this analysis, we have previously documented an increased risk of readmission for heart failure within 30 days after discharge among subjects discharged home on ACE-I in combination with aspirin, compared to those discharged on ACE-I without aspirin [16]. The lack of an adverse mortality effect from combined ACE-I and aspirin use might seem contradictory to increased early readmissions from the same combination. However, factors that affect heart failure readmission rates might not necessarily be operative in causing increased mortality. Further, it is also possible that the negative interaction between ACE-I and aspirin is most obvious in the early period after hospitalization, when patients are most susceptible to adverse hemodynamic interactions, and that it wanes with time.
4.1. Implications Experimental data are suggestive of the negative effects of aspirin coadministration on hemodynamic, ventilatory, and endothelial benefits provided by angiotensin-converting enzyme inhibition in heart failure [7,8]. On the other hand, observational clinical evidence, based on our study and posthoc analyses of SOLVD and CONSENSUS II studies, indicates that combined therapy with ACE-I and aspirin is not any
213
worse than ACE-I therapy without aspirin. A ‘negative interaction’ between antiplatelet agents and enalapril is nevertheless apparent in the SOLVD and CONSENSUS II studies, since randomization to enalapril (versus placebo) led to a survival benefit in nonusers of antiplatelet agents, but not so in users of antiplatelet agents in both these studies. However, in the current era of unquestioned benefit from ACE-I in heart failure [2], the clinically relevant question is whether addition of aspirin, in patients already receiving ACE-I, is beneficial, neutral, or detrimental to long-term mortality. Our results indicate that among patients on chronic ACE-I therapy, the addition of aspirin does not have any adverse effect on long-term mortality. Further, among patients with CAD, a trend towards improvement in mortality is seen in patients treated with the combination of ACE-I and aspirin, compared to those treated with ACE-I without aspirin. This finding is suggestive that combined aspirin and ACE-I use should be strongly considered in patients with symptomatic heart failure and CAD.
4.2. Limitations Our study suffers from the limitations inherent to nonrandomized, retrospective analyses. The dose or duration of therapy with either ACE-I or aspirin, or other agents known to be beneficial in heart failure, are not known. The study period preceded the publication of large-scale clinical studies on the beneficial effects of b-blocker therapy in symptomatic heart failure. As such, b-blocker therapy was generally avoided in patients with symptomatic heart failure; this may limit the validity of our results in the current era of more generous b-blocker use. Clinical follow-up was available in only 94.4% of the cohort. Furthermore, the number of patients in our study is relatively small, particularly in analyses of subgroups. Nevertheless, our results indicate that combined therapy with ACE-I and aspirin is not worse than ACE-I without aspirin.
Acknowledgements We are indebted to Mario Vaz for his assistance in the preparation of this manuscript.
214
K. J. Harjai et al. / International Journal of Cardiology 88 (2003) 207–214
References [1] Applefeld MM. Chronic congestive heart failure. Where have we been? Where are we heading. Am J Med 1986;80(Suppl 2B):73–7. [2] Garg R, Yusuf S. Overview of randomized trials of angiotensinconverting enzyme inhibitors on mortality and morbidity in patients with heart failure. J Am Med Assoc 1995;273:1450–6. [3] Second International Study of Infarct Survival (ISIS-2) Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both or neither among 17 187 cases of suspected acute myocardial infarction. Lancet 1988;2((8607)):349–60. [4] Hennekens CH, Buring JE, Sandercock P, Collins R, Peto R. Aspirin and other antiplatelet agents in the secondary and primary prevention of cardiovascular disease. Circulation 1989;80:749–56. [5] The Steering Committee of the Physicians’ Health Study Research Group. Final report on the aspirin component of the ongoing Physicians’ Health Study. New Engl J Med 1989;321:129–35. [6] Hall D, Zeitler H, Rudolph W. Counteraction of the vasodilator effects of enalapril by aspirin in severe heart failure. J Am Coll Cardiol 1992;20(7):1549–55. [7] Guazzi M, Marenzi G, Alimento M, Contini M, Agostoni P. Improvement of alveolar–capillary membrane diffusing capacity with enalapril in chronic heart failure and counteracting effect of aspirin. Circulation 1997;95:1930–6. [8] Oates JA. Antagonism of antihypertensive drug therapy by nonsteroidal anti-inflammatory drugs. Hypertension 1988;11:114–6, Review. [9] Houston MC. Nonsteroidal anti-inflammatory drugs and antihypertensives. Am J Med 1991;90:42S–7S.
[10] Johnson AG, Nguyen TV, Day RO. Do nonsteroidal anti-inflammatory drugs affect blood pressure? A meta-analysis. Ann Intern Med 1994;121:289–300. [11] Riegger GA, Kahles HW, Elsner D, Kromer EP, Kochsiek K. Effects of acetylsalicylic acid on renal function in patients with chronic heart failure. Am J Med 1991;90:571–5. [12] Harjai KJ, Nunez E, Turgut T, Shah MP, Humphrey JS, Newman J, Cheirif J, Smart FW, Ventura HO. The independent effects of left ventricular ejection fraction on short-term outcomes and resource utilization following hospitalization for heart failure. Clin Cardiol 1999;22:184–90. [13] Amico AF, Lichtenberg GS, Reisner SA, Stone CK, Schwartz RG, Meltzer RS. Superiority of visual versus computerized echocardiographic estimation of radionuclide left ventricular ejection fraction. Am Heart J 1989;118:1259–65. [14] Al-Khadra AS, Salem DN, Rand WM, Udelson JE, Smith JJ, Konstam MA. Antiplatelet agents and survival: a cohort analysis from the studies of left ventricular dysfunction (SOLVD) trial. J Am Coll Cardiol 1998;31:419–25. [15] Nguyen KN, Aursnes I, Kjekshus J. Interaction between enalapril and aspirin on mortality after acute myocardial infarction: subgroup analysis of the Cooperative New Scandinavian Enalapril Survival Study II. Am J Cardiol 1997;79:115–9. [16] Harjai KJ, Nunez E, Turgut T, Newman T. Effect of combined aspirin and angiotensin-converting enzyme inhibitor therapy versus angiotensin-converting enzyme inhibitor therapy alone on readmission rates in heart failure. Am J Cardiol 2001 15;87(4):483–7, A7.