Extent of and reasons for nonuse of implantable cardioverter defibrillator devices in clinical practice among eligible patients with left ventricular systolic dysfunction.

BACKGROUND Several studies that used claims and registry data have reported that 40% to 80% of patients eligible for an implantable cardioverter defibrillator (ICD) fail to receive one in clinical practice, and the rates are especially high among women and blacks. The extent and documented reasons for nonuse of ICDs among patients with left ventricular systolic dysfunction are unknown. METHODS AND RESULTS Using hospital claims and clinical data, we identified patients hospitalized with a heart failure diagnosis and left ventricular ejection fraction ≤30% between January 1, 2007, and August 30, 2007, at a tertiary-care center. Using claims data, we determined placement of an ICD or cardiac resynchronization therapy with defibrillation device at any time up to 1 year after hospitalization. Medical records for patients without an ICD were abstracted to determine reasons for nonuse. Patients with an ICD were compared with patients without an ICD and also with patients without an ICD who did not have any contraindication for an ICD as identified through chart abstraction. Of the 542 potentially eligible patients identified, 224 (41%) did not have an ICD. In the initial adjusted analysis, female sex (odds ratio=1.90; 95% CI, 1.28 to 2.81) and increasing age (odds ratio=1.07; 95% CI, 1.04 to 1.11) were associated with a higher likelihood of not having an ICD. After detailed chart review, of the 224 patients without an ICD, 117 (52%) were ineligible for the device and 38 (17%) patients refused the device, resulting in only 69 (13%) patients eligible for an ICD who failed to receive one. In this subsequent adjusted analysis, remaining factors associated with a higher likelihood of not having an ICD were absence of ventricular arrhythmias (odds ratio=4.93; 95% CI, 2.56 to 9.50), noncardiology hospital service (odds ratio=3.73; 95% CI, 1.98 to 7.04), and lack of health insurance (odds ratio=3.10; 95% CI, 1.48 to 6.46). CONCLUSIONS On the basis of a detailed chart review, the true rate of ICD underuse may be substantially lower than previous estimates. In addition, after accounting for ICD eligibility criteria, patient sex and age disparities in ICD therapy were no longer present.