National Patterns of Use and Effectiveness of Angiotensin-Converting Enzyme Inhibitors in Older Patients With Heart Failure and Left Ventricular Systolic Dysfunction
Background— Although ACE inhibitors are underprescribed for heart failure, factors associated with their use are not well described. Furthermore, the effectiveness of ACE inhibitors has been questioned in some populations, potentially contributing to underuse. Our objectives were to assess the correlates of ACE inhibitor use and the relationship between ACE inhibitor prescription and mortality in older patients with heart failure.
Methods and Results— We studied a national sample aged ≥65 years who had survived hospitalization for heart failure between April 1998 and March 1999 or July 2000 and June 2001, restricting the analysis to patients with left ventricular systolic dysfunction and without a documented contraindication to use of ACE inhibitors (n=17 456). Factors associated with ACE inhibitor prescription at discharge and the relationship between ACE inhibitor prescription and death within 1 year were assessed with hierarchical logistic models. Secondary analyses assessed therapeutic substitution with angiotensin receptor blockers (ARBs). ACE inhibitors were prescribed to only 68% of this ideal cohort, and 76% received either an ACE inhibitor or an ARB. Patient, physician, and hospital factors were weak predictors of prescription, except for serum creatinine (RR for 133 to 221 μmol/L=0.87, 95% CI 0.85 to 0.89; RR for ≥222 μmol/L=0.53, 95% CI 0.49 to 0.57 compared with ≤132 μmol/L). ACE inhibitor prescription was associated with lower crude 1-year mortality (33.0% versus 42.1%, P<0.001), lower risk of death after adjustment (RR 0.86, 95% CI 0.82 to 0.90), and lower mortality regardless of patient gender, age, race, or serum creatinine level.
Conclusions— ACE inhibitors were widely underprescribed despite evidence of a favorable impact on survival in a broad range of patients with heart failure. These results emphasize the importance of ongoing efforts to translate clinical trial results into practice.
Received April 14, 2004; revision received June 7, 2004; accepted June 15, 2004.
Randomized trials have established the efficacy of ACE inhibitors in treating heart failure and left ventricular systolic dysfunction (LVSD).1,2 Despite this evidence and endorsements in clinical practice guidelines,3 ACE inhibitor use in appropriate patients has not been optimized.4–6 A recent national study of 22 quality indicators for 6 conditions among Medicare beneficiaries reported that ACE inhibitors are underused in ideal candidates hospitalized with heart failure and LVSD and that prescription rates have not increased recently.7 The specific patient and provider characteristics associated with ACE inhibitor underuse, however, are not well known. Because of the importance of understanding this persistent gap in heart failure care, the identification of factors associated with ACE inhibitor prescribing is particularly relevant.
See p 644
Shortfalls in practice may reflect uncertainty about the effectiveness of ACE inhibitors in some heart failure patients. Many patients seen in clinical practice, including the elderly and those with renal insufficiency, have not been well represented in clinical trials.8,9 Post hoc analyses of clinical trials have also raised questions about the benefits of ACE inhibitor therapy in some groups, including black patients and women.10,11 Because further randomized trials comparing ACE inhibitors with placebo are not likely forthcoming, an assessment of ACE inhibitor effectiveness in a community-based sample is important to validate recommendations to treat all patients regardless of age, race, gender, or renal function.
The study of ACE inhibitor use and effectiveness in heart failure is complicated by the emergence of angiotensin receptor blockers (ARBs) as accepted treatment for some patients. Although current guidelines do not endorse the substitution of ARBs for patients who tolerate ACE inhibitors,3 this practice may account for some of the perceived gaps in the quality of heart failure care.
To address these issues, we evaluated factors associated with the prescription of ACE inhibitors, the relationship between ACE inhibitor prescription and mortality, and the importance of ARB substitution in a national sample of hospitalized heart failure patients with LVSD without treatment contraindications. This study was intended to identify targets for initiatives to enhance ACE inhibitor prescription and to quantify ACE inhibitor effectiveness in a community-based sample of patients with heart failure.
The National Heart Care Project
The National Heart Care (NHC) Project is a Centers for Medicare & Medicaid Services initiative designed to improve the quality of care for Medicare beneficiaries hospitalized with heart failure.12 Fee-for-service beneficiaries hospitalized with a principal discharge diagnosis of heart failure (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes 402.01, 402.11, 402.91, 404.01, 404.91, or 428) between April 1998 and March 1999 or July 2000 and June 2001 were identified. In each period, discharged patients were sorted by age, gender, race, and treating hospital, and up to 800 per state were selected randomly. All records were included if fewer than 800 hospitalizations occurred in a state during a sampling period. Medical records were reviewed by trained abstractors. Data quality was ensured through the use of abstraction software and random record reabstraction. Patients with invalid social security numbers, those receiving long-term hemodialysis, those transferred to another hospital, and those leaving against medical advice were excluded. The NHC sample consisted of 78 882 records (39 477 were from 1998 to 1999, and 39 405 were from 2000 to 2001).
The sample for the present study was an ideal cohort (ie, composed entirely of patients for whom ACE inhibitors are appropriate). All candidates for inclusion had documented LVSD (left ventricular ejection fraction <0.40 or a qualitative description of moderate or severe LVSD) based on the evaluation closest to the discharge date (n=24 258). Assessments with echocardiography, left ventricular angiography, or radionuclide ventriculography were accepted. Patients younger than 65 years (n=2441), those who died during hospitalization (n=1330), and those who were discharged to hospice care (n=558) were excluded. If a patient appeared more than once (n=1232), only 1 record was randomly selected. Patients with documented contraindications to ACE inhibitor treatment (n=1486), including bilateral renal artery stenosis, serum potassium nearest to discharge >5.5 mmol/L, enrollment in a randomized clinical trial of an ACE inhibitor alternative, or any documented reason for withholding treatment (eg, severe aortic stenosis or prior adverse reaction such as cough or angioedema), were also excluded. A total of 6802 records of patients with LVSD were thus excluded, which resulted in a cohort of 17 456.
Additional Data Sources
Attending physician characteristics were ascertained by linkage with the American Medical Association Physician Masterfile,13 and hospital characteristics were determined with American Hospital Association annual surveys.14,15 Patient vital status within 1 year after discharge was determined with the Medicare Enrollment Database.16
Patients were evaluated for the prescription of an ACE inhibitor at discharge. Differences in prescription rates in strata defined by patient clinical characteristics, physician factors, and hospital characteristics were compared with χ2 tests.
Hierarchical logistic regression was used to identify factors independently associated with prescription of an ACE inhibitor. Patient, physician, and hospital variables with a univariate probability value <0.1 and prespecified variables of clinical significance (including demographic characteristics, diabetes, coronary artery disease, hypertension, and serum creatinine) were included. Subsequently, individual variables were eliminated by backward selection with a retention criterion of <0.05. To account for therapeutic substitution of ARBs for an ACE inhibitor, secondary analyses were repeated to identify correlates of the prescription of either an ACE inhibitor or an ARB. The discrimination of these models was assessed with the c-index.
Crude 1-year mortality rates for patients who were and were not prescribed ACE inhibitors were compared with χ2 tests. The independence of this relationship was assessed in hierarchical multivariable logistic models. Candidates for inclusion in these models included patient, physician, and hospital characteristics and other discharge medications, including β-blockers and aspirin. After the full model was constructed, a variable indicating the discharge prescription of an ACE inhibitor was included. In secondary analyses, the prescription of either an ACE inhibitor or ARB was used.
To assess for heterogeneity in the relationship between ACE inhibitors and 1-year mortality in clinically important subgroups, stratified multivariable models were constructed. The statistical significance of differences in these relationships was assessed with the probability value for interaction.
All analyses used probability weights proportional to the inverse sampling fraction of the state. ORs were converted to estimated risk ratios (RRs).17 Statistical analyses were conducted with Stata 7.0 (Stata Corporation), SAS version 8.0 (SAS Institute), and MLWiN (Institute of Education, London). Analyses of the NHC data were approved by the Institutional Review Board of the University of Colorado.
The mean age of the cohort was 78 years, with nearly 20% being older than 84 years. Cardiovascular comorbidities and other coexisting conditions were common (Table 1⇓). More than one third had a serum creatinine level ≥133 μmol/L, and 61% had severe LVSD (left ventricular ejection fraction <0.30 or qualitatively severe). The majority of patients were treated by generalist physicians and were hospitalized in not-for-profit hospitals.
Patterns of ACE Inhibitor Prescription
Among the study cohort, 68% were prescribed an ACE inhibitor at hospital discharge. Prescription rates did not change significantly between 1998 and 1999 (69%) or 2000 and 2001 (67%; P=0.07). Prescription rates were higher among patients treated with β-blockers and aspirin and were highest in the Northeast US Census region. ACE inhibitor prescription rates did not exceed 74% in any group (Table 1⇑).
Patient and provider characteristics were modestly correlated with ACE inhibitor prescription at hospital discharge, except for serum creatinine at discharge (compared with creatinine ≤132 μmol/L, RR for creatinine 133 to 221 μmol/L=0.87, 95% CI 0.85 to 0.89; RR for creatinine ≥222 μmol/L=0.53, 95% CI 0.49 to 0.57; Table 2). The c-index for the full model was 0.71.
ACE Inhibitor Prescription and Mortality
Crude mortality at 1 year was lower among patients who received an ACE inhibitor than among those who did not (33.0% versus 42.1%, P<0.0001; unadjusted RR 0.78, 95% CI 0.75 to 0.81). ACE inhibitor prescription at discharge was associated with a 14% relative reduction in 1-year mortality (RR 0.86, 95% CI 0.82 to 0.90) after multivariable adjustment and was associated with a lower adjusted risk of death in almost all clinical subgroups (Figure). Although ACE inhibitor prescription was significantly associated with lower mortality in patients of both genders, the association was stronger in men (RR 0.80, 95% CI 0.74 to 0.86) than in women (RR 0.93, 95% CI 0.86 to 1.00, P for interaction=0.005). The relationship also differed significantly according to the presence and severity of coronary artery disease (P for interaction=0.01). The relationship between ACE inhibitor prescription and mortality was otherwise comparable when tested across the strata defined by other patient characteristics, but with trends toward a stronger relationship in the oldest patients (P for interaction=0.07), patients without diabetes (P=0.06), those with higher serum creatinine (P=0.06), and those discharged with instructions to take aspirin (P=0.06).
ARB Substitution Patterns and Mortality
ARBs were prescribed at discharge to 1726 patients (10%). Prescription rates for either ACE inhibitors or ARBs were comparable between sampling periods and did not exceed 82% in any group (Table 1⇑). Patient and provider factors were only modestly associated with the prescription of ACE inhibitors or ARBs, except for serum creatinine (Table 2). The c-index for the full model was 0.76.
Discharge prescription of either ACE inhibitors or ARBs was associated with a 17% relative reduction in mortality after adjustment (RR 0.83, 95% CI 0.79 to 0.88). In addition to similar interactions by gender and coronary artery disease, treatment with an ACE inhibitor or ARB in patients with serum creatinine levels ≤132 μmol/L was associated with a significantly different reduction in mortality (RR 0.88, 95% CI 0.81 to 0.96) compared with those with serum creatinine levels of 133 to 221 μmol/L (RR 0.73, 95% CI 0.66 to 0.80) or ≥222 μmol/L (RR 0.65, 95% CI 0.51 to 0.80, P for interaction=0.002).
This national study of a community-based cohort of older patients with LVSD and without treatment contraindications demonstrates that ACE inhibitors have been widely underprescribed. Therapeutic substitution with ARBs did not fully account for shortfalls in ACE inhibitor utilization. Significantly improved survival associated with ACE inhibitor prescription was seen in a wide range of patients, although the magnitude of this relationship differed in some subgroups. These findings suggest that the beneficial effects of ACE inhibitors reported in randomized trial populations are applicable to elderly community-based patients.
The underprescription of ACE inhibitors for patients in this ideal cohort was pervasive. This finding is discouraging in light of the long-standing availability of strong clinical evidence,1,2 expert consensus,18 guidelines endorsing the use of ACE inhibitors,3 and the resources dedicated to improving the quality of heart failure care. Although suboptimal ACE inhibitor use in ideal candidates has been demonstrated previously, these studies have been limited by small sample sizes, narrowly defined populations, or relatively old data.4–6 More recently, national data from the NHC project provided an assessment of contemporary care patterns but did not assess the factors associated with ACE inhibitor use.7
There are several possible explanations for the underuse of ACE inhibitors. The present study demonstrates that ARB substitution does not completely explain the treatment gap. Clinicians may be reluctant to use ACE inhibitors because of concerns of increased risks of adverse events in some patients (eg, in patients with renal dysfunction)19,20 or because of the lack of data from clinical trials in other groups. In some cases, systems may not be structured to ensure that patients are receiving optimal care. Finally, inadequate clinician awareness of the potential benefits of ACE inhibitors may play a role in some cases.
Although randomized trials have established the importance of ACE inhibitors in many patients with LVSD, these trials typically excluded patients with renal dysfunction or patients who were of advanced age. For example, the Studies of Left Ventricular Dysfunction (SOLVD) treatment trial excluded patients with a serum creatinine level in excess of 221 μmol/L or those who were older than 80 years of age.2 The patterns of ACE inhibitor use in the present study suggest the possibility of therapeutic nihilism in patients with renal insufficiency, whereas the mortality analyses indicate that older patients and those with renal insufficiency may have the most to gain from treatment.
Pooled analyses of randomized trials have also called into question the efficacy of ACE inhibitors in other populations. A post hoc analysis of the SOLVD trials found no reduction in hospitalization rates or mortality with enalapril in black patients.10 A recent meta-analysis of clinical trials could not demonstrate evidence of reduced mortality with ACE inhibitor treatment in some women with heart failure.11 Although these studies did not conclude that treatment of heart failure should necessarily differ by race or gender, there has been some concern that such studies may be mistakenly construed as adequate evidence of the lack of efficacy of ACE inhibitors in these subgroups.
Although relationships in observational studies may reflect unmeasured confounding, including the effects of differences in care after hospitalization, the strength of the associations, the consistency of the results across subgroups, and the concordance with previous randomized trial data support the validity of these findings. The 14% relative reduction in mortality associated with ACE inhibitors demonstrated in the present study is consistent with the 20% odds reduction in a previous meta-analysis of treatment trials (OR 0.80, 95% CI 0.74 to 0.87).21 The present study, however, similar to those of aspirin and β-blockers after myocardial infarction, provides evidence of the effectiveness of an intervention when data outside the environment of controlled trials had been lacking.22,23
Certain issues should be considered in interpreting these results. As a study of hospitalized fee-for-service Medicare patients, these results may not apply to all heart failure patients. However, the present study focused on a large population for whom data supporting the use of ACE inhibitors are limited. Second, we could not assess adherence in follow-up, to identify patients starting treatment after discharge or to assess the effects of ACE inhibitor dosing. Biases due to the misclassification of ACE inhibitor use, however, are likely to result in an underestimation of the benefits. Third, the apparent underuse of ACE inhibitors may in some cases reflect undocumented contraindications. However, ARB substitution, which is appropriate for many patients with contraindications to ACE inhibitors, did not account for the entire gap in prescription rates and resulted in similar results in the mortality analyses.
In conclusion, ACE inhibitors were commonly underused in older patients with LVSD and without drug-specific contraindications. ACE inhibitor prescription was independently associated with reduced mortality in a wide range of patients, including women, black patients, the very old, and those with renal insufficiency. These findings support existing guideline recommendations for widespread ACE inhibitor use in heart failure patients with LVSD and highlight the ongoing need for efforts to optimize the use of ACE inhibitors with the goal of ensuring that all eligible patients benefit from these life-prolonging medications.
The analyses on which this publication is based were performed under contract No. 500-02-CO 01 entitled “Utilization and Quality Improvement Organization for the State of Colorado” sponsored by the Centers for Medicare & Medicaid Services, Department of Health and Human Services. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government. The authors assume full responsibility for the accuracy and completeness of the ideas presented. This article is a direct result of the Health Care Quality Improvement Program initiated by the Centers for Medicare & Medicaid Services, which has encouraged identification of quality-improvement projects derived from analysis of patterns of care, and therefore required no special funding on the part of this contractor. Ideas and contributions to the author concerning experiences in engaging with the issues presented are welcomed. Dr Masoudi has received honoraria from and served on the speakers’ bureau of AstraZeneca and Pfizer. Dr Havranek has received honoraria from and served as a consultant to Bristol-Myers Squibb. Dr Foody has received honoraria from and served on the speakers’ bureau of Pfizer, Bristol-Myers Squibb, and Merck.
Dr Masoudi is supported by NIH/NIA research career award K08-AG01011. Dr Foody is supported by NIH/NIA research career award K08-AG20623 and NIA/Hartford Foundation fellowship in geriatrics. Mr Rathore is supported by NIH/National Institute of General Medical Sciences medical scientist training grant GM07205.
Consulting Editor for this article was Burton E. Sobel.
Hunt SA, Baker DW, Chin MH, et al. ACC/AHA guidelines for the evaluation and management of chronic heart failure in the adult. Circulation. 2001; 104: 2996–3007.
American Hospital Association. The AHA Annual Survey Database: Fiscal Year 1998 Documentation. Chicago, Ill: Health Forum, American Hospital Association; 1999.
American Hospital Association. The AHA Annual Survey Database: Fiscal Year 2000 Documentation. Chicago, Ill: Health Forum, American Hospital Association; 2001.
Krumholz HM, Radford MJ, Ellerbeck EF, et al. Aspirin in the treatment of acute myocardial infarction in elderly Medicare beneficiaries. Circulation. 1995; 92: 2841–2847.