Background: Based on current patient selection criteria, a significant proportion of recipients of cardiac resynchronization therapy (CRT) do not respond to treatment. The purpose of this analysis is to identify predictors and characterize the timing of response to CRT in patients with advanced heart failure.
Methods: Patients randomized to receive CRT in the MIRACLE and MIRACLE-ICD trials, designed to assess the benefit of CRT compared with standard medical therapy in patients with advanced heart failure, left ventricular ejection fraction <0.35, and QRS > or =130 milliseconds, were included for this analysis. Patients with an improvement of > or =1 New York Heart Association (NYHA) class from baseline to the 6-month follow-up were considered responders and those who had no change or worse NYHA class or died were classified as nonresponders. Responders were subdivided into early (within 1-3 months) and late (6 months).
Results: One hundred forty-three (64%) of 224 and 190 (61%) of 313 patients in the MIRACLE and MIRACLE-ICD trials, respectively, responded to therapy, with 81 (57%) of 143 and 100 (53%) of 190 responding early. Several but differing factors predicted CRT response and timing in the two trials with a high sensitivity (89%-90%) but, owing to a low specificity (31%-49%), a modest predictive accuracy (66%-75%).
Conclusions: Based on improvement of > or =1 NYHA class, less than two thirds of patients enrolled in the MIRACLE or MIRACLE-ICD trials responded to CRT, with just more than half responding within the first month. Several factors predicted CRT response and timing, but given their modest predictive accuracy, comparable for both studies, additional methods for selecting candidates most likely to benefit from CRT are needed.