Add CABG to Meds for Quality-of-Life Gains in Ischemic HF

September 21, 2011

September 21, 2011 (Boston, Massachusetts) — An array of quality-of-life measures improved significantly more for patients in the CABG arm than in the medical-therapy-only arm of the controversial Surgical Treatment for Ischemic Heart Failure (STICH) trial, according to researchers here at the Heart Failure Society of America (HFSA) 2011 Scientific Meeting [1].

Seen consistently over three years, those significant gains with CABG seemed less open to debate than the trial's primary findings for the two treatment arms, interpretations of which have famously varied, as heartwire described when STICH was presented at the ACC meeting in April.

As reported then, survival over an average of five years wasn't significantly altered by adding CABG to medical therapy in the trial. But in alternate "as-treated" analyses and with various statistical adjustments, some point out, mortality did fall significantly for those who underwent surgery compared with the conservative approach, and it also had the edge for such secondary end points as cardiovascular death and death/cardiovascular hospitalization.

For many patients with heart failure, as Dr Paul Heidenreich (Stanford University, CA), not a STICH investigator, noted here at the HFSA sessions, quality of life takes precedence over greater survival as a treatment goal. For them, he observed, the trial's quality-of-life analysis, presented by Dr Daniel B Mark (Duke Clinical Research Institute, Durham, NC) and which Heidenreich was critiquing as discussant, may be more meaningful than the primary results.

Dr Daniel B Mark

From 2002 to 2007 at 99 centers in 22 countries, STICH randomized 1212 patients with an LVEF <35% and heart failure who were considered candidates for CABG to receive or not receive the surgery on top of medical therapy. Over a median of 56 months, the trial's primary end point of all-cause mortality was 36% in the CABG/meds group and 41% for the meds-only patients (p=0.12). The advantage for CABG patients edged into statistical significance only after controlling for baseline features (p=0.039).

Secondary clinical end points also showed advantages for the surgery group, including a 19% drop in risk of cardiovascular death (p<0.05) and a 26% fall in the composite of all-cause death or cardiovascular hospitalization (p<0.0001).

The patients also prospectively completed a battery of quality-of-life tests at four months and at one year, two years, and three years after randomization. It consisted of 23 queries about physical limitation, symptoms, social limitation, and self-efficacy from the widely used Kansas City Cardiomyopathy Questionnaire (KCCQ), which were scored from 0 to 100, with a higher score indicating more favorable conditions.

Important to interpreting the results, Mark said in his presentation, is the distinction between statistical and clinical significance. A statistically significant difference was also considered clinically significant if it was different by at least 5 KCCQ points.

Value is going to be highly sensitive to how patients rate these improvements in quality of life. If they rate them high enough, even these small differences may prove that bypass is of value

In general, after four months and at 36 months, he noted, quality-of-life scores tended to be significantly better in the CABG group but not clinically better, with the divergence broadening at both 12 and 24 months such that the CABG gains were doubly significant. In an exception, CABG significantly improved angina frequency both ways from baseline out to two years but in neither way thereafter.

Mark said that in analyses that have yet to be reported, differences in NYHA functional class and Canadian Cardiovascular Society angina class apparently follow a similar pattern.

Differences in Selected KCCQ Scores Between CABG+Meds and Meds-Alone Groups, Over Time (All Differences Favor CABG+Meds)

Parameter 4 mo 12 mo 24 mo 36 mo
Summary score +2.6 +5.3 +4.5 +2.7
p 0.006 <0.001 0.001 0.04
Quality-of-life satisfaction score +4.9 +7.6 +5.5 +2.7
p <0.001 <0.001 <0.001 0.05
Physical limitation +1.6 +4.1 +4.3 +2.6
p 0.05 0.003 0.009 0.08
Angina frequency +6.5 +6.7 +3.9 +1.3
p <0.001 <0.001 <0.001 0.19

KCCQ scores range from 0 to 100; scores consistently were not significantly different between the groups at baseline. Higher score corresponds to more favorable status for each domain

Adding CABG to meds also showed significant gains when the measure was percentage of patients who showed a clinically significant >5-point improvement in KCCQ summary scores,

Proportion of STICH Patients Showing Gain of >5 KCCQ Points at Each Follow-Up, by Treatment Group

Time from randomization (mo) CABG+meds, n=610 (%) Meds only, n=602 (%) p
4 63 55 0.007
12 69 58 0.0004
24 67 57 0.002
36 66 58 0.03

"Are these results applicable to the general population?" asked Heidenreich in his commentary. "I'm going to conclude yes, but we need to consider [whether there was] a placebo effect." Importantly, conventional blinding to treatment assignment was impossible, a problem with many randomized trials of therapy that includes surgery.

And as Heidenreich, Mark, and questioners in the audience observed, the randomization process itself was subject to selection bias, complicating interpretation of the results.

"It’s a dilemma that we face in all of our trials, and we haven't figured out how to get past it, but it is an important issue," Mark said.

Dr Eugene Braunwald (Brigham and Women's Hospital, Boston, MA), comoderator for the session featuring Mark's presentation, responded, "Well, we face it more in some trials than in others. I would say that in STICH it was a bigger problem than in for example, ASCEND."

Dr Randall C Starling

Outcomes at 180 days in the Acute Study of Clinical Effectiveness of Nesiritide in Decompensated Heart Failure (ASCEND-HF) trial, which tested the addition of nesiritide (Natrecor, Scios/Johnson & Johnson) to standard treatment in acute decompensated heart failure, had been presented earlier in the same afternoon session by Dr Randall C Starling (Cleveland Clinic, OH) [2]. Those late outcomes, he reported, mirrored the primary outcomes at 30 days, in which nesiritide seemed safe for the kidneys but had no effect, good or bad, on mortality, as covered by heartwire .

Heidenreich presented a quick, informal cost-effectiveness analysis of STICH (Mark noted that the trial's formal analysis has yet to be completed) to see whether the "small but real" quality-of-life gains with CABG "have value" according to traditional cost-effectiveness criteria.

He concluded that those gains "may just not be large enough in magnitude or duration to justify the increased cost of the procedure--with the one caveat that value is going to be highly sensitive to how patients rate these improvements in quality of life. If they rate them high enough, even these small differences may prove that bypass is of value."

STICH was supported by the National Heart, Lung, and Blood Institute. Mark discloses consulting for Aventis and receiving research grants from Eli Lilly and Medtronic. Heidenreich and Braunwald had no disclosures. Starling reports receiving research grants from Novartis and Thoratec, receiving honoraria for speaking from Medtronic, and consulting or serving on an advisory board for Novartis and Medtronic.