The responsibility of managing atrial fibrillation (AF) and atrial flutter (AFL) falls predominantly on the Medicare system. Patients with AF or AFL often have a range of cardiovascular (CV) co-morbidities and are frequently hospitalized for AF and other CV causes. The present retrospective cohort study used medical claims data to evaluate the rates of hospitalization and inpatient mortality in elderly (aged ≥65 years) patients with AF or AFL with Medicare supplemental insurance. The data were extracted from the United States Thomson Reuters MarketScan Medicare Supplemental and Coordination of Benefits Database (January 2004 to December 2007). Patients aged ≥65 years with ≥1 inpatient or ≥2 outpatient nondiagnostic claims for AF or AFL and ≥12 months of continuous enrollment before their index AF or AFL diagnoses were identified. The frequencies of hospitalization and inpatient death were evaluated over the postindex study period (mean 24.3 months). Of an eligible study population of 55,774 patients with AF or AFL (mean age 77.9 years, 52.2% men), 28,939 patients (51.9%) were hospitalized (all causes) with nonfatal outcomes, 12,652 (22.7%) were rehospitalized, and 1,592 (2.9%) died in the hospital. Higher proportions of patients were hospitalized for non-CV than for CV causes (35.6% vs 27.2%). For CV hospitalizations culminating in inpatient death (n = 516), the most common admission diagnoses were major bleeding, stroke or transient ischemic attack, and congestive heart failure. In conclusion, elderly patients with AF or AFL undergo frequent hospitalization for CV and non-CV causes. Measures that lower inpatient admission rates, particularly readmission rates, may reduce the increasing cost of treating patients with AF or AFL with Medicare supplemental insurance.
The prevalence of atrial fibrillation (AF) and atrial flutter (AFL) is substantially higher in elderly patients than in the population as a whole. Therefore, much of the cost burden of managing these conditions falls on the Medicare system. One estimate suggested that the overall prevalence of AF in Medicare patients increased from 3.2% in 1992 to 6.0% in 2002. The cost of managing AF is substantial, largely driven by inpatient costs, and is projected to increase as the population ages. However, there is limited detailed information on the management of Medicare patients with AF or AFL or on the rates of hospitalization, rehospitalization, and inpatient mortality. The present study was designed to provide a real-world evaluation of the burden of hospitalization in a large group of patients aged ≥65 years with AF or AFL who are covered by Medicare supplemental insurance.
Methods
This retrospective observational cohort study used data from January 1, 2004, to December 31, 2007, from the United States Thomson Reuters MarketScan Medicare Supplemental and Coordination of Benefits Database. This database contains information on eligible retirees (aged ≥65 years) with Medicare supplemental insurance plans. It includes all employer and Medicare coordination of benefits. The United States Thomson Reuters MarketScan databases have been used for numerous health economic and outcomes research studies, which have formed the basis for >275 peer-reviewed reports. Because all patient data were deidentified in compliance with the Health Insurance Portability and Accountability Act of 1996, the study was exempt from institutional review board approval.
Patients with ≥1 inpatient or ≥2 outpatient nondiagnostic medical claims (on different days) with International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM), diagnosis codes of 427.31 (AF) or 427.32 (AFL) were identified. Claims had to occur between January 1, 2005, and December 31, 2005. The first qualifying AF or AFL diagnosis was designated the index diagnosis. Eligible patients were aged ≥65 years and had ≥12 months of continuous enrollment and prescription data availability before and after the index AF or AFL diagnosis. Patients who experienced inpatient death within 12 months of the index AF or AFL diagnosis were exempted from the postindex continuous enrollment criterion.
Patients were excluded if they had histories of heart failure (HF), which was defined as ≥1 inpatient or outpatient nondiagnostic (i.e., radiology, laboratory, or pathology) claim with an ICD-9-CM diagnosis code for HF (428.xx, 398.91, 402.01, 402.11, 402.91, 404.01, 404.03, 404.11, 404.13, 404.91, or 404.93) before the index AF or AFL diagnosis. Patients were also excluded if they had evidence of transient AF or AFL associated with cardiac surgery (a single AF diagnosis within 30 days of surgery) or thyroid disease (ICD-9-CM diagnosis code for hyperthyroidism or an outpatient prescription claim for methimazole or propylthiouracil in the 12 months before the index diagnosis). Preexisting AF was defined as the presence of an AF claim ≤12 months before the index diagnosis.
Patient demographic characteristics were assessed at the date of the index diagnosis (index date). The proportions of patients with co-morbidities were assessed over the 12-month period preceding the index date. The presence of the following cardiovascular (CV) co-morbidities was assessed: hypertension, structural heart disease, coronary artery disease, diabetes, valvular heart disease, pulmonary disease, ischemic stroke, peripheral vascular disease, thyroid disease, and cardiomyopathies. The Charlson co-morbidity index and chronic disease score were used to provide aggregate measures of co-morbidities, and the cumulative risk for stroke was assessed using the CHADS 2 (congestive HF, hypertension, age >75 years, diabetes, and previous stroke) scoring system. The use of medications and nonpharmacologic treatments (e.g., radiofrequency catheter ablation) was assessed during the 12-month period preceding the index diagnosis of AF or AFL.
Hospitalizations and inpatient deaths were classified using the primary ICD-9-CM diagnosis code and categorized on the basis of the main causes of CV hospitalization used in A Placebo-Controlled, Double-Blind, Parallel Arm Trial to Assess the Efficacy of Dronedarone 400 mg bid for the Prevention of Cardiovascular Hospitalization or Death From Any Cause in Patients With Atrial Fibrillation/Atrial Flutter (ATHENA). An additional category of “other CV causes” was included to capture CV events not included in ATHENA but that met the criteria for CV hospitalizations defined by the Institute for Clinical Evaluative Sciences. This included rheumatic fever and rheumatic heart disease, post–myocardial infarction syndrome, acute cor pulmonale, chronic pulmonary heart disease, pericardial or endocardial disease, cardiomyopathy, conduction disorders, paroxysmal tachycardia, premature beats, unspecified cardiac dysrhythmia, systolic or diastolic HF, atherosclerosis, aortic aneurysm and dissection, arterial embolism and thrombosis, polyarteritis nodosa and allied conditions, and diseases of the capillaries, veins and lymphatics.
Findings are presented separately for the entire follow-up period and the first 12 months of follow-up to identify temporal changes in the patterns of hospitalization and mortality and to enable comparison with previous studies.
Results
A total of 55,774 Medicare patients aged ≥65 years with nontransient AF or AFL and no histories of HF were identified from the database population ( Figure 1 ). This represents about 50% of all the patients with AF identified from the Medicare database. The mean postindex follow-up period was 24.3 ± 8.1 months. The mean age was 77.7 years, and 52.2% were men ( Table 1 ). In the 12 months before their index AF or AFL diagnoses, 86.4% of patients had received treatment with antiarrhythmic, anticoagulant (68% of patients, with warfarin most widely used), or rate-control drugs or any combination ( Table 2 ). Nonpharmacologic interventions were relatively uncommon, with <1% of patients having claims for radiofrequency catheter ablation.
| Characteristic | Value |
|---|---|
| Age (years) | 77.7 ± 6.8 |
| Age category (years) | |
| 65–69 | 7,533 (14%) |
| 70–74 | 11,390 (20%) |
| 75–79 | 14,256 (26%) |
| 80–84 | 13,236 (24%) |
| ≥85 | 9,359 (17%) |
| Gender | |
| Male | 29,129 (52.2%) |
| Female | 26,645 (47.8%) |
| Region | |
| North central | 19,514 (35.0%) |
| Northeast | 5,740 (10.3%) |
| South | 15,989 (28.7%) |
| West | 14,450 (25.9%) |
| Unknown | 81 (0.1%) |
| Charlson co-morbidity index | 1.1 ± 1.5 |
| Chronic disease score | 7.3 ± 3.4 |
| CHADS 2 score | 1.8 ± 0.9 |
| 0 | 3,216 (5.8%) |
| 1 | 17,022 (30.5%) |
| 2 | 26,455 (47.4%) |
| 3 | 6,104 (10.9%) |
| 4 | 2,426 (4.3%) |
| 5 | 551 (1.0%) |
| 6 | 0 (0%) |
| Co-morbidities | |
| Hypertension | 44,877 (80.5%) |
| Structural heart disease | 18,354 (32.9%) |
| Coronary artery disease | 12,871 (23.1%) |
| Diabetes | 10,603 (19.0%) |
| Valvular heart disease | 7,773 (13.9%) |
| Pulmonary disease | 6,012 (10.8%) |
| Ischemic stroke | 4,186 (7.5%) |
| Peripheral vascular disease | 2,575 (4.6%) |
| Thyroid disease | 1,969 (3.5%) |
| Cardiomyopathies | 998 (1.8%) |
| Medication | n (%) |
|---|---|
| Antiarrhythmic agents | |
| Dofetilide | 220 (0.4%) |
| Sotalol | 3,065 (5.5%) |
| Procainamide | 71 (0.1%) |
| Propafenone | 1,446 (2.6%) |
| Flecainide | 1,118 (2.0%) |
| Amiodarone | 5,630 (10.1%) |
| Disopyramide | 222 (0.4%) |
| Other antiarrhythmic agents | 230 (0.4%) |
| Anticoagulant agents | |
| Warfarin | 32,184 (57.7%) |
| Clopidogrel | 4,673 (8.5%) |
| Warfarin-clopidogrel combination | 981 (1.8%) |
| Rate-control agents | |
| Acebutolol | 133 (0.2%) |
| Atenolol | 8,057 (14.4%) |
| Bisoprolol | 577 (1.0%) |
| Carvedilol | 1,863 (3.3%) |
| Metoprolol | 14,688 (26.3%) |
| Propranolol | 952 (1.7%) |
| Other β blockers | 531 (0.1%) |
| Any β blocker | 25,823 (46.3%) |
| Diltiazem | 8,317 (14.9%) |
| Verapamil | 2,804 (5.0%) |
| Other calcium channel blockers | 9,941 (17.8%) |
| Digoxin | 16,733 (30.0%) |
| Other CV medications | |
| Diuretics | 23,994 (43.0%) |
| Angiotensin-converting enzyme inhibitors | 18,184 (32.6%) |
| Thyroid replacement therapy | 10,838 (19.4%) |
| β-adrenergic stimulant | 7,204 (12.9%) |
| Nonpharmacologic treatments | |
| Radiofrequency catheter ablation | 375 (0.7%) |
| Atrioventricular node ablation | 226 (0.4%) |
| Pacemaker insertion | 1,641 (2.9%) |
| Surgical maze procedure | 216 (0.4%) |
| Electrical conversion | 1,230 (2.2%) |
Over the entire study period, 28,939 patients were hospitalized with nonfatal outcomes (all causes), and 12,652 were rehospitalized ( Figure 2 ). Most of the recorded hospitalizations occurred during the first year, with 38.3% of patients hospitalized during this period. Patients had a mean of 0.9 ± 1.2 hospitalizations (all cause) over the entire follow-up period, in comparison with 0.4 ± 0.7 over the 1 year. Of the 28,939 patients who were hospitalized, 12,652 (43.7%) were readmitted. Higher proportions of patients over the entire follow-up period and the first year were hospitalized for non-CV versus CV causes ( Figures 2 and 3 ). The mean duration of all-cause hospitalizations was 5.5 ± 8.5 days when assessed over the first year and the entire follow-up period. Hospitalizations for CV causes had a mean duration of 4.4 ± 5.6 days over the 2 follow-up periods.
Among the prespecified list of CV causes for hospitalization, a higher proportion of patients were hospitalized with primary diagnoses of AF or other supraventricular arrhythmias than with any other CV diagnosis, over the first year and the entire follow-up period ( Figure 4 ). Other common primary diagnoses included congestive HF (4.4% of patients over the entire follow-up period and a total of 2,859 hospitalizations; 14.1% of CV hospitalizations), coronary atherosclerosis (3.2% and 1,966), and transient ischemic attack or stroke (3.7% and 2,234) ( Figure 4 ).
