Limited data exist describing the differences in the medical treatment of patients with heart failure with preserved ejection fraction (HF-PEF) from those with heart failure with reduced ejection fraction (HF-REF) in more generalizable population-based cohorts. We studied patients with incident HF diagnosed from 2005 to 2008 from 4 sites participating in the Cardiovascular Research Network. These patients, their medication profile, and left ventricular systolic function status were identified from the hospital discharge and ambulatory visit diagnoses, pharmacy dispensing information, and imaging reports found in the health plan electronic databases and through chart review. The study population consisted of 6,210 patients with newly diagnosed HF-PEF and 3,914 patients with newly diagnosed HF-REF. The mean age of our study population was 73 years, 48% were women, and 74% were white. The patients with HF-REF were less likely to have been treated with various cardiac and HF-related medications before their index HF event; however, they were significantly more likely to have been treated with new cardiac medications and HF therapies after the diagnosis of HF than were the patients with HF-PEF. After controlling for several potentially confounding factors, the patients with HF-PEF were significantly less likely to have been treated with multiple cardiac drug regimens (adjusted odds ratio 0.69, 95% confidence interval 0.59 to 0.81) and multiple HF-related therapies (adjusted odds ratio 0.40, 95% confidence interval 0.38 to 0.42) than were patients with HF-REF. In conclusion, the present results from a large, population-based sample suggest considerable variation in the previous and new use of different cardiac medication classes of drugs in patients with HF-PEF versus HF-REF.
Although heart failure is recognized as a contemporary epidemic in the United States, the epidemiology and treatment of patients with heart failure (HF) and preserved left ventricular ejection fraction (HF-PEF) is less well known and described. Although a number of therapies are available to improve the symptoms associated with chronic HF with reduced ejection fraction (HF-REF) and its long-term prognosis, no specific guidelines are available for the treatment of patients with HF-PEF. The contemporary treatment of these patients is also relatively unknown. Using data from the Cardiovascular Research Network (CVRN), we examined the differences in the medical treatment of patients with incident HF-REF compared to that of those with newly diagnosed HF-PEF.
Methods
The source population for the present study included adult members from 4 participating health plans within the National Heart, Lung, and Blood Institute–sponsored CVRN. The study sites included Kaiser Permanente Northern California, Kaiser Permanente Colorado, Kaiser Permanente Northwest, and the Fallon Community Health Plan. These sites were identified on the basis of providing care to a socioeconomically diverse population across varying clinical practice settings and geographic locations. The institutional review boards at the participating sites approved the present study.
The CVRN virtual data warehouse at each study site served as the primary data source for the present investigation. The CVRN virtual data warehouse is a distributed standardized data resource composed of electronic data sets at each CVRN site. The virtual data warehouse includes the demographic, pharmacy, outpatient and inpatient laboratory test results, and healthcare usage data for those patients in the CVRN health plans.
The study sample consisted of adults aged ≥21 years with HF who were enrolled in the participating health plans from January 2005 to December 2008. The diagnosis of HF was determined by hospitalization with a primary discharge diagnosis of HF and/or having ≥3 ambulatory visits coded for HF with ≥1 visit with a cardiologist. The following International Classification of Diseases, 9th edition, codes were used to identify the patients with HF: 398.91, 402.01, 402.11, 402.91, 404.01, 404.03, 404.11, 404.13, 404.91, 404.93, 428.0, 428.1, 428.20, 428.21, 428.22, 428.23, 428.30, 428.31, 428.32, 428.33, 428.40, 428.41, 428.42, 428.43, and 428.9. We have previously shown that the vast majority (>95%) of patients with a primary discharge diagnosis of HF using these codes were confirmed as having clinical HF by chart review using the Framingham clinical criteria.
Information on the quantitative and/or qualitative assessments of left ventricular systolic function was abstracted from the results of the echocardiograms, nuclear imaging studies, and left ventriculography from the health plan databases, complemented by manual review of the electronic health records. We defined PEF as either a left ventricular ejection fraction of ≥50% and/or a physician’s qualitative determination of normal ejection fraction or PEF . A REF was defined as a left ventricular ejection fraction of ≤40% and/or a physician’s qualitative determination of moderately or severely REF. Patients (n = 1,870) with borderline ejection fraction findings (>40% to <50%) were excluded from the study sample because we wanted to focus the present report on patients with PEF versus REF findings. We restricted our study sample to patients with an incident episode of HF during a 5-year “look-back” period. We determined whether the patient had been previously diagnosed with HF before their index date of HF from the ambulatory and hospitalization data sources.
We ascertained information on coexisting illnesses according to the medical diagnoses or receipt of selected procedures using the relevant International Classification of Diseases, 9th edition, codes, laboratory test results, or filled outpatient prescriptions from health plan hospitalization discharge, ambulatory visit, laboratory, and pharmacy databases. We collected the baseline and follow-up data on the diagnoses of selected co-morbidities and interventional procedures according to the relevant International Classification of Diseases, 9th edition, and/or Common Procedural Terminology codes. We also ascertained the available ambulatory results for blood pressure, serum lipid levels, estimated glomerular filtration rate, and serum hemoglobin levels on or before the index date of diagnosis of preserved or reduced systolic function.
Prescription data were used to identify previous (≤120 days before, but not including, the index date of diagnosis of HF and remaining active within 30 days of the index date) and postdiagnosis (≤90 days after the index date of diagnosis) use of cardiac and HF medications. The cardiac medications examined included anticoagulants, antiplatelet agents, calcium channel blockers, thiazide diuretics, statins, and other lipid-lowering drugs. The HF-related medications included angiotensin-converting enzyme (ACE) inhibitors/angiotensin receptor blockers (ARBs), aldosterone antagonists, β blockers, digoxin, loop diuretics, nitrates, and hydralazine. We further subdivided the HF therapies into those that have been shown to affect the prognosis (ACE inhibitors/ARBs, aldosterone antagonists, β blockers, nitrates, and hydralazine) and those used to treat the symptoms (digoxin, loop diuretics).
We characterized the present cohort of patients with HF-REF and HF-PEF with regard to several sociodemographic, medical history, and clinical characteristics. We modeled the differences, through the use of logistic regression modeling, in the new receipt of each of our different medication categories (cardiac medication only, HF medication only, HF prognosis-related therapy, and HF symptom-related therapy) between newly diagnosed patients with preserved versus reduced systolic function, simultaneously controlling for age, gender, race/ethnicity, history of cardiovascular disease, receipt of coronary reperfusion/revascularization procedures, dyslipidemia, hypertension, diabetes, hospitalized bleeding episodes, diagnosed dementia or depression, chronic liver or lung disease, systemic cancer, estimated glomerular filtration rate, systolic blood pressure, hemoglobin, serum lipid values, and study site.
Results
Of the 10,124 patients with newly diagnosed HF from 2005 to 2008 at the 4 study sites, 6,210 patients (61.3%) had HF-PEF and 3,914 patients had HF-REF. The mean age of our study population was approximately 73 years, 48% were women, and 74% were white. The mean ± SD ejection fraction in patients with HF-PEF and HF-REF was 57.2 ± 19.7% and 29.1 ± 9.6%, respectively.
The patients with incident HF-PEF were significantly older and included a greater proportion of women, whites, patients with a history of coronary heart disease, cerebrovascular disease, atrial fibrillation, peripheral arterial disease, previous bleeding episodes, and other selected co-morbidities than patients with HF-REF ( Table 1 ). These patients had lower baseline estimated glomerular filtration rate and serum hemoglobin levels, greater systolic but lower diastolic blood pressure levels, greater high-density lipoprotein cholesterol levels, and lower serum low-density lipoprotein cholesterol levels.
| Variable | PEF (n = 6,210) | REF (n = 3,914) | p Value |
|---|---|---|---|
| Age (yrs) | 74.7 ± 12.1 | 69.1 ± 14.0 | <0.001 |
| Women | 3,543 (57.1%) | 1,277 (32.6%) | <0.001 |
| Race | <0.001 | ||
| White | 4,759 (76.6%) | 2,742 (70.1%) | |
| Black | 428 (6.9%) | 387 (9.9%) | |
| Medical history (n) | 0.51 | ||
| Acute myocardial infraction/unstable angina | 599 (9.6%) | 362 (9.2%) | |
| Coronary artery bypass surgery and/or percutaneous coronary intervention | 610 (9.8%) | 368 (9.4%) | 0.49 |
| Ischemic stroke or transient ischemic attack | 438 (7.1%) | 187 (4.8%) | <0.001 |
| Atrial fibrillation or flutter | 2,195 (35.3%) | 918 (23.5%) | <0.001 |
| Mitral and/or aortic valvular disease | 1,315 (21.2%) | 408 (10.4%) | <0.001 |
| Peripheral arterial disease | 398 (6.4%) | 191 (4.9%) | <0.001 |
| Implantable cardioverter defibrillator | 22 (0.4%) | 71 (1.8%) | <0.001 |
| Pacemaker | 282 (4.5%) | 150 (3.8%) | 0.09 |
| Dyslipidemia | 3,833 (61.7%) | 2,151 (55.0%) | <0.001 |
| Hypertension | 5,114 (82.4%) | 2,478 (63.3%) | <0.001 |
| Diabetes mellitus | 1,217 (19.6%) | 701 (17.9%) | 0.03 |
| Hospitalized bleeding episodes | 306 (4.9%) | 114 (2.9%) | <0.001 |
| Dementia | 480 (7.7%) | 195 (5.0%) | <0.001 |
| Depression | 1,141 (18.4%) | 524 (13.4%) | <0.001 |
| Chronic lung disease | 2,369 (38.1%) | 1,150 (29.4%) | <0.001 |
| Chronic liver disease | 242 (3.9%) | 129 (3.3%) | 0.12 |
| Systemic cancer | 493 (7.9%) | 249 (6.4%) | 0.003 |
| Estimated glomerular filtration rate (ml/min/1.73 m 2 ) | 61.1 ± 24.7% | 66.4 ± 23.8% | <0.001 |
| Hemoglobin (g/dl) | 12.5 ± 2.0% | 13.5 ± 1.9% | <0.001 |
| Systolic blood pressure (mm Hg) | 138.6 ± 24.3% | 132.8 ± 21.9% | <0.001 |
| Diastolic blood pressure (mm Hg) | 76.8 ± 13.6% | 80.3 ± 14.3% | <0.001 |
| High-density lipoprotein (g/dl) | 49.0 ± 15.5% | 47.2 ± 14.6% | <0.001 |
| Low-density lipoprotein (g/dl) | 97.4 ± 33.9% | 102.1 ± 35.5% | <0.001 |
Patients with newly diagnosed HF-PEF were significantly more likely to have been previously treated with ACE inhibitors/ARBs, anticoagulants, β blockers, calcium channel blockers, digoxin, diuretics, nitrates, hydralazine, and statins than patients with HF-REF. Patients with HF-REF were significantly, but only slightly, more likely to have been previously treated with aldosterone antagonists ( Table 2 ).
| Variable | PEF (n = 6,210) | REF (n = 3,914) | p Value |
|---|---|---|---|
| Medication | |||
| Angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers | 3,080 (49.6%) | 1,681 (42.9%) | <0.001 |
| Aldosterone receptor agonist | 93 (1.5%) | 99 (2.5%) | <0.001 |
| Anticoagulants | 1,198 (19.3%) | 513 (13.1%) | <0.001 |
| Antiplatelets | 484 (7.8%) | 287 (7.3%) | 0.42 |
| β Blocker | 3,877 (62.4%) | 1,539 (39.5%) | <0.001 |
| Calcium channel blocker | 2,099 (33.8%) | 599 (15.3%) | <0.001 |
| Digoxin | 498 (8.1%) | 250 (6.4%) | <0.005 |
| Diuretics (loop) | 2,085 (33.6%) | 973 (24.9%) | <0.001 |
| Diuretics (thiazide) | 1,728 (27.8%) | 678 (17.3%) | <0.001 |
| Nitrates | 723 (11.6%) | 408 (10.4%) | 0.06 |
| Nitrates and hydralazine | 71 (1.1%) | 14 (0.4%) | <0.001 |
| Statin | 2,980 (48.0%) | 1,572 (40.2%) | <0.001 |
∗ Receiving medication ≤120 days before, but not including, index date of HF diagnosis and remaining active within 30 days of index date of diagnosis.
In terms of our 4 major cardiac and HF treatment categories, patients with HF-PEF were significantly more likely than patients with HF-REF to have been treated with all cardiac-related drugs examined (76.3% vs 59.4%), HF-related drugs (82.3% vs 64.7%), HF medications shown to improve prognosis (76.4% vs 57.5%), and medications shown to manage HF-related symptoms (38.0% vs 28.5%) before the initial HF diagnosis (p <0.001 for all).
At their HF diagnosis, and ≤90 days thereafter, a significantly greater proportion of patients with newly diagnosed HF-REF were treated with each of the cardiac- and HF-related medications examined compared to those with newly diagnosed HF-PEF ( Table 3 ). Patients with HF-REF were significantly more likely than those with HF-PEF to have been newly treated with selected cardiac (58.8% vs 50.3%) and HF-related medications (92.8% vs 81.8%, p <0.001). These differences were also noted when we examined the prescribing of HF medications shown to improve a patient’s prognosis (89.3% vs 65.1%, p <0.001). However, the HF drugs used to provide symptomatic relief were prescribed similarly in our primary comparison groups (75.4% vs 73.6%).
| Medication | PEF (n = 6,210) | REF (n = 3,914) | p Value |
|---|---|---|---|
| Angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers | 1,343 (42.9) | 1,732 (77.6) | <0.001 |
| Aldosterone receptor agonist | 315 (5.2) | 828 (21.7) | <0.001 |
| Anticoagulants | 1,139 (48.8) | 1,886 (79.4) | <0.05 |
| Antiplatelets | 516 (9.0) | 830 (22.7) | <0.001 |
| β Blocker | 2,965 (71.9) | 2,090 (71.1) | <0.001 |
| Calcium channel blocker | 326 (7.3) | 199 (6.2) | <0.001 |
| Digoxin | 809 (14.7) | 759 (21.7) | <0.001 |
| Diuretics (loop) | 242 (3.9) | 218 (5.6) | 0.45 |
| Diuretics (thiazide) | 767 (23.8) | 966 (41.3) | <0.05 |
| Nitrates | 183 (3.2) | 196 (5.4) | <0.001 |
| Nitrates and hydralazine | 796 (15.9) | 601 (17.7) | <0.001 |
| Statins | 565 (13.7) | 201 (6.1) | <0.001 |
∗ Receiving selected medication at HF diagnosis and ≤90 days thereafter.
After controlling for a variety of demographic, medical history, and clinical characteristics, the patients with newly diagnosed HF-PEF were significantly less likely to have been prescribed selected cardiac medications (adjusted odds ratio [OR] 0.77; 95% confidence interval [CI] 0.66 to 0.91), HF-related medications (adjusted OR 0.42; 95% CI 0.34 to 0.52), HF therapies shown to improve prognosis (adjusted OR 0.26; 95% CI 0.24 to 0.29), and HF therapies used to treat HF-related symptoms (adjusted OR 0.82; 95% CI 0.77 to 0.88) than patients with HF-REF after their index diagnosis.
Of the various cardiac medications examined before the patient’s index HF diagnosis, 16.4% were not treated with any of these medications, 37.4% of all patients were treated with any 1 therapy, 29.4% were treated with any 2, and 16.8% were treated with any ≥3 of these medications. Multiple cardiac medication use (≥2) was more common in patients with HF-PEF than with HF-REF ( Table 4 ). Of the new cardiac medication users, little to no differences were found in the use of multiple cardiac medications in patients with HF-PEF versus HF-REF ( Table 4 ).
