Running Into Snags with the STICH Trial: The Sharp Pain One Investigator Feels is the Misinterpretation of Data | CardioSource WorldNews Interventions

ACCEL | The Surgical Treatment for Ischemic Heart Failure (STICH) trial is famous for its surprising result suggesting less value than expected for coronary artery bypass graft (CABG) surgery versus optimal medical therapy. At the outset, it was expected to demonstrate that CABG plus medical therapy would improve mortality over medical therapy alone.

While it's remembered for its failure to show a clear advantage for CABG surgery, Raymond J. Gibbons, MD, from the Mayo Clinic in Rochester, Minnesota, thinks the STICH trial is widely misunderstood and misquoted.

In STICH, investigators enrolled 1,212 patients with a left ventricular ejection fraction (LVEF) <35% and coronary artery disease (CAD) amenable to CABG who were then randomized to medical therapy plus CABG (n = 610) or medical therapy alone (n = 602).1 The primary outcome was all-cause mortality, while major secondary outcomes included death from cardiovascular causes and death from any cause or hospitalization for cardiovascular causes. Although investigators found no significant difference in all-cause mortality (p = 0.12), significantly fewer CABG patients experienced cardiovascular deaths (28% vs. 33%; p = 0.05) or death from any cause or cardiovascular hospitalization (58% vs. 68%; p < 0.001).

Initially, the design of STICH required viability testing with single-photon emission computed tomography (SPECT) but when this proved too much of an obstacle to enrollment, a revised protocol allowed for optional viability testing and permitted use of either SPECT or dobutamine echocardiography. Researchers were trying to determine whether such assessment could help identify which patients with CAD and LV dysfunction would gain the greatest survival from the addition of CABG to medical management.

While viability testing was strongly encouraged, it was performed in only 51% of the patients, usually after randomization. Among those who underwent viability testing, medical therapy plus CABG was reported for 298 patients versus 303 who received medical therapy alone. Of the patients undergoing viability testing, reported in a separate substudy, 487 had viable myocardium and 114 did not.2

Importantly, those with viable myocardium at baseline had significantly better LVEF, LV end-diastolic and end-systolic volume indices, and lower rates of previous MIs.

During follow-up of the 601 patients (median 5.1 years), there were significantly fewer deaths in patients with myocardial viability (37% vs. 51%; p = 0.003). However, after adjusting for other baseline prognostic variables, viability status was not significantly associated with mortality (p = 0.21). In terms of secondary endpoints, patients with myocardial viability had significantly lower rates of mortality (p = 0.003) and composite of death or cardiovascular hospitalization (p < 0.001), but only the latter remained significant on multivariable analysis.

When overlaying the treatment strategy per patient group, patients receiving CABG in addition to medical therapy had lower rates in regard to the endpoints, but the difference did not reach statistical significance, whether considering death (p = 0.53), cardiovascular death (p = 0.70), or death/ cardiovascular hospitalization (p = 0.39).

Trial Widely Misunderstood
What often gets missed in recalling the study are the words of the paper's own authors: "The lack of an unequivocally significant between-group difference in the outcome of death from any given cause was more likely to have been due to limited power and limited duration of follow-up than to a true lack of benefit of CABG." According to Dr. Gibbons, this point appears to have been lost among many of those who continue to discuss the findings of the trial.

One of the big problems with the viability substudy, said Dr. Gibbons, is the breadth of important information that was not in the paper proper but rather reported in the supplemental material. Since Mayo Clinic was a study site, Dr. Gibbons is well aware that four different SPECT protocols were allowed and the same 50% threshold for relative tracer activity was used for both sestamibi and thallium. "I have significant concerns about that threshold because on the basis of previously published studies from our laboratory, it tends to underestimate the amount of infarction," Dr. Gibbons said. "[SPECT and dobutamine echo] are really very different measures of viability and, to some degree, they are comparing apples and oranges."

In fact, one of the supplemental tables from the paper reveals that viability assessed with SPECT alone was associated with mortality (p = 0.007), but dobutamine echo alone was not (p = 0.277). While this was seen only with univariate analysis and not multivariate analysis, Dr. Gibbons pointed out that the significant difference reported for mortality based on viability was driven by the SPECT data.

There was also a big difference between patients in each group in terms of mortality according to viability and treatment assignment: 487 patients with viability versus only 114 without. It's even worse when evaluating effect of therapy on those without viability: only 60 patients were on medical therapy, and 54 underwent CABG.

Finally, Dr. Gibbons summarized other limitations of the viability substudy:

  • Patient selection was very different between those with and without viability.
  • Only a small percentage (19%) of patients was without viability.
  • The high rate of baseline medical therapy helps explain why these results differ from previous studies where baseline medical therapy was either not very good or not even reported.
  • More than one-quarter of patients enrolled had single-vessel CAD.
  • No magnetic resonance imaging or positron emission tomography was permitted.

The authors of the overall trial also acknowledged that their study was underpowered. Dr. Gibbons agrees, saying that the STICH study illustrates that absence of evidence of benefit is not evidence of lack of benefit—particularly for CABG.


  1. Bonow RO, Maurer G, Lee KL, et al. N Engl J Med. 2011;364:1617-25.
  2. Velazquez EJ, Lee KL, Deja MA, et al. N Engl J Med. 2011;364:1607-16.

To listen to an interview with Raymond J. Gibbons, MD, about nuclear cardiology and PET, visit The interview was conducted by Dipti Itchhaporia, MD.

Keywords: CardioSource WorldNews Interventions

< Back to Listings