The preliminary results of the COMPANION study were first presented at the 2003 American College of Cardiology Scientific Sessions and previously reported by heartwire.

The study was a parallel, randomized clinical trial in 1600 patients with moderate or severe heart failure with QRS >120 ms and P-R interval >150 ms. Patients were randomized in a 1:2:2 fashion to optimal medical therapy (including beta blockers, diuretics, ACE inhibitors or ARBs, spironolactone, and/or digoxin); optimal drug therapy plus CRT; or optimal drug therapy plus CRT with an ICD (CRT-D).

In reporting the final COMPANION data, Bristow told the audience that the primary end point of the study, a combination of all-cause death and all-cause hospitalizations over 12 months, was significantly reduced by 20% for both the CRT and CRT-D arms of the study. CRT alone was also associated with a nonsignificant trend toward a 24.0% reduction in all-cause mortality, a secondary end point of the study, while CRT-D increased the mortality reduction to 36.0%, a highly significant result. The reduction in mortality in the CRT-D arm was down from the 43% seen in the preliminary results.

In an interview with heartwire, Bristow discussed some of the observed differences in ischemic and nonischemic HF patients. He noted that from the standpoint of ICDs, based on MADIT II criteria, the devices are indicated only for ischemic HF. However, it is thought that nonischemic HF patients could also benefit.

"There has been this sense that has developed that it's important for ischemic cardiomyopathy patients to get ICDs, that there is something special happening here, but the heart failure people haven't necessarily bought into that, because people with nonischemic cardiomyopathy also die suddenly," said Bristow.

When the COMPANION investigators analyzed the all-cause mortality data in the CRT-D arm, one of the more important findings of the study, due its magnitude, according to Bristow, was that they observed a 50% reduction in mortality in the nonischemic patients compared with a 27% reduction in the ischemic HF patients. While Bristow said the degree of reduction is not statistically different by test of interaction, the overall 50% reduction in the nonischemic patients is statistically significant.

"The end of the story is that the addition of an ICD to CRT makes a difference in the nonischemic patients," said Bristow. He added, "It looks like, based on the COMPANION subgroup analyses between CRT and CRT-D, that the ICD is important to obtain the mortality reduction in nonischemic patients."

The investigators also found significant differences in systolic blood pressurea surrogate marker of LV contractilityin patients receiving CRT, both with and without the ICD, compared with the optimal-medical therapy arm. Systolic blood pressure was improved in both device groups, while diastolic blood pressure was unchanged, indicating improved systolic function of the heart chamber, said Bristow.

CRT and CRT-D patients on a beta blocker and spironolactone also fared better than patients not taking the drugs, said Bristow, who stated that "CRT appears to work in the face of the best medical therapy we have."

Dr John G Cleland

Commenting on the results for heartwire, Dr John G Cleland (University of Hull, Kingston-Upon-Hull, UK) said the COMPANION results continue to impress.

Regarding the secondary-end-point mortality data, he remarked the COMPANION mortality benefits appeared to shrink a little, down from 43% to 36%, in the CRT-D arm but were maintained in the CRT-alone patient group.

"We've got tantalizing evidence of a mortality benefit from cardiac resynchronization, but we've haven't been able to securely prove it," said Cleland. "However, the overall take on this study is definitely positive."

Cleland said that while there exists a positive mortality meta-analysis from CRT, the results exclude failed CRT trials.

Cleland added he is the principal investigator of the large Cardiac Resynchronization in Heart Failure (CARE-HF) clinical trial in Europe. That trial is evaluating mortality and morbidity among moderate to severe heart failure patients who receive CRT. The primary end point is all-cause mortality or CV hospitalizations. More than 800 patients from 80 centers in Europe are currently enrolled, said Cleland.

"I would imagine that with COMPANION, CARE-HF, and other existing data, by this time next year we should have a very secure position on whether or not there is a mortality/morbidity benefit with cardiac resynchronization," said Cleland. "I personally don't think we're home and dry yet, but we're getting there."