To the Editor: Chronic heart failure (HF) is a significant problem in the United States, affecting 5.7 million Americans. Studies have consistently demonstrated substantial racial and ethnic disparities in the use of guideline-recommended device therapies related to HF care. For example, although blacks have the highest age-adjusted rates of sudden cardiac arrest (1), they remain significantly less likely than their white counterparts to receive an implantable cardioverter-defibrillator (ICD) (2). Cardiac resynchronization therapy with defibrillation (CRT-D) represents another important therapeutic option for those with HF. The use of CRT-D can significantly reduce rehospitalizations and all-cause mortality while improving quality of life and functional status (3); however, blacks and Hispanics have historically been less likely to receive CRT-D than whites (4). By using the National Cardiovascular Data Registry's ICD Registry, we compared trends in CRT-D use among potentially eligible patients receiving an ICD to determine whether racial and ethnic disparities have changed over time.

We queried the ICD Registry for patients eligible for CRT-D on the basis of New York Heart Association functional class III or IV symptoms, left ventricular ejection fraction (LVEF) ≤35%, native QRS width ≥120 ms, and receipt of optimal medical therapy between April 2006 and March 2010. Optimal medical therapy was defined as treatment with a beta-blocker and either an angiotensin-converting enzyme inhibitor or an angiotensin receptor blocker. Individual sites were responsible for establishing and submitting the race and ethnicity of patients receiving a device. We excluded patients who had a previous ICD or who required epicardial lead placement, who did not meet the criteria for the CRT-D use, or whose ICD type was unknown. We examined the use of CRT-D over different time periods and tested the trend using the Cochran–Armitage test among the subgroups determined by race. Interactions between race and time with the use of CRT-D were examined using a hierarchical logistic model to see whether the time trends are different among different race groups.

A final study population of 107,096 patients was identified from cases entered between 2006 and 2010. Of these, 87,692 (81.9%) received CRT-D. In the overall study population, the mean age was 70.7 years, 70.1% of the patients were male, and 83.8% were white. The mean LVEF was 23.7%. In the overall study population, the use of CRT-D in eligible patients significantly increased from 80.4% in the first year to 84.0% in the fourth year (p < 0.001) (Table 1). Among eligible whites, blacks, and Hispanics, there was a significant increase in CRT-D use over the study period. After adjusting for age, race, gender, atrial fibrillation/atrial flutter, cerebrovascular disease, chronic lung disease, diabetes, ischemic heart disease, duration of symptom since initial HF onset, prior hospitalizations for HF, previous myocardial infarction, LVEF, QRS duration, creatinine, sodium, brain natriuretic peptide, hospital owner, hospital region, electrophysiologist operator ICD training, physician volume, and the patients' clustering among hospitals in the hierarchical model, blacks and Hispanics remained less likely to receive CRT-D compared with whites (black vs. white odds ratio: 0.69; 95% confidence interval: 0.65 to 0.73, p < 0.001; Hispanic vs. white odds ratio: 0.84; 95% confidence interval: 0.78 to 0.91; p < 0.001). After adjustment, the effect of time on CRT-D use in eligible patients did not significantly vary according to the level of race/ethnicity (p = 0.68). In a sensitivity analysis of racial and ethnic subgroups restricted to Medicare patients only, the temporal trends among each subgroup persisted (whites: p < 0.001, blacks: p = 0.02), although these were not significant among Hispanics (p = 0.22).

Underlying racial and ethnic differences in treatment patterns have been shown to contribute to the overall underuse of CRT-D (4), as is true for other cardiovascular technologies. Similar disparities have been shown for other interventions in cardiovascular care, including coronary artery bypass grafting, cardiac catheterization, and ICD use (5). The use of these technologies has increased over time, but underlying treatment gaps have persisted across racial and ethnic lines. Before this study, it was unknown whether CRT-D, a more recent technology to emerge with widespread applications, has diffused unevenly over time as preceding technologies have. Our analysis of 107,096 patients in the ICD Registry indicates that although adherence to this recommendation has improved over time, the widespread application of CRT-D in eligible patients is hampered by the same underuse in racial and ethnic subgroups that has hindered the diffusion of other cardiovascular technologies.

To improve the quality of care for patients with advanced HF, further initiatives are needed that broaden the use of CRT-D in eligible patients overall, but particularly those belonging to racial and ethnic subgroups. Several factors that account for these differences (i.e., informed and noninformed personal preferences, provider biases, geographic variations, and access to care) need to be addressed. Increasing awareness among patients and providers about the benefits of CRT-D is an important mechanism by which the underuse in appropriate patients can be reduced. Likewise, tools that enable health care providers to systematically identify eligible patients at the point of care can improve implementation. Last, quality improvement programs and performance measures that help practices benchmark their quality can help reduce differences in care. Among such initiatives, the American College of Cardiology has established the Coalition to Reduce Racial and Ethnic Disparities in Cardiovascular Outcomes to address these unchanged trends in care.

It is important to note that although the Centers for Medicare & Medicaid Services mandates participation in the ICD Registry for hospitals performing ICD and CRT-D implantations, data are not available on those patients who might be eligible but have not been referred for or received an ICD. Likewise, those patients who were offered therapy, but declined, are not included. Also, data are not available to identify cases in which CRT-D was planned but may have been aborted because of technical limitations; regardless, these data would not have been affected by patient race or ethnicity.

The use of CRT-D seems to be improving in the overall population and in racial and ethnic subgroups; however, disparities in the use of CRT-D have remained over time. Future efforts to improve the broad and systematic use of CRT-D in eligible patients are needed to minimize differences in care and improve the treatment of patients with advanced HF.