Observational Studies Stir Controversy in Heart Failure

November 27, 2012

DURHAM, North Carolina and STOCKHOLM, Sweden — Two new observational studies in heart failure are provoking some controversy, as they both conflict with conclusions reached from previous randomized trials.

One study suggests aldosterone antagonists are not associated with improved survival or hospitalizations in elderly heart-failure patients [1], and the other study suggests ACE inhibitors/angiotensin-receptor blockers (ARBs) do benefit heart-failure patients with preserved ejection fraction [2].

Both studies are published in the November 28, 2012 issue of the Journal of the American Medical Association.

McMurray Troubled by JAMA Publication

Asked by heartwire how to interpret the results of these new studies, Dr John McMurray (University of Glasgow, Scotland), who wasn't involved in either analysis, was adamant that they should not be used to question results of previous randomized trials. "Frankly, I'm troubled that leading journals such as JAMA publish studies like this. I think they confuse doctors and patients and do more harm than good. We all know that you simply cannot deduce the effect of treatment from nonrandomized comparisons like this, no matter how much statistical adjustment you try to do," he commented.

What is it that JAMA is trying to achieve by publishing these studies?

McMurray noted that there are three large prospective randomized trials with aldosterone-receptor antagonists, each of which showed a reduction in mortality, and there are no conflicting findings. "Conversely, we have two large randomized trials showing no benefit of ARBs in patients with heart failure and preserved ejection fraction and no additional trial data to suggest otherwise."

He further pointed out that observational studies often mistakenly suggest treatment benefit that doesn't exist. Two examples include the hypotheses that statins reduce all-cause mortality in patients with heart failure and that hormone-replacement therapy reduces cardiovascular events in postmenopausal women, both of which were suggested in observational studies but then disproved in prospective randomized trials.

He added; "What is it that JAMA is trying to achieve by publishing these studies? Don't they think they will encourage physicians to avoid beneficial treatments and even use treatments known not to be beneficial?"

JAMA declined to comment on the issue for heartwire .

Aldosterone-Antagonists Uncertainty?

In the current paper on aldosterone antagonists, Dr Adrian Hernandez (Duke Clinical Research Institute, Durham, NC) and colleagues say that although these agents have been shown beneficial in randomized trials, their uptake has been slow in clinical practice, possibly due to uncertainty about their effectiveness and safety outside clinical trials, which is especially relevant for patients at high risk of hyperkalemia.

For their study, Hernandez et al used clinical registry data linked to Medicare claims from 2005 to 2010 to examine outcomes of 5887 elderly hospitalized patients with heart failure and reduced ejection fraction, including Cox proportional hazards models and adjustments for selection bias.

Results showed that aldosterone-antagonist use was associated with a decrease in hospitalizations related to heart failure (HR 0.87) but was not associated with improved overall survival (HR 1.04) or reduced cardiovascular readmissions (HR 1.00) at three years. In addition, hyperkalemia was more common with aldosterone-antagonist use (HR 1.50).

They conclude: "Strict protocols for careful monitoring and early follow-up after initiation of aldosterone-antagonist therapy are needed. Additional research is needed to evaluate the clinical effectiveness of aldosterone antagonists in the broad population of patients with heart failure and to identify strategies to overcome disparities between findings of clinical efficacy and clinical effectiveness."

In an accompanying editorial [3], however, Dr James Fang (University Hospitals Case Medical Center, Cleveland, OH) points out that despite rigorous attempts to adjust for confounding and selection bias, important differences between treatment groups persisted, thereby limiting the robustness of the findings.

ACE/ARBs in Preserved Ejection Fraction

In the ACE/ARB study in heart failure with preserved ejection fraction, Dr Lars Lund (Karolinska Institutet, Stockholm, Sweden) and colleagues used the Swedish Heart Failure registry to identify 16 216 patients with heart failure with preserved ejection fraction (>40%). Of these, 12 543 were treated and 3673 were not treated with renin-angiotensin-system (RAS) antagonists. In two analyses--one using a cohort matched 1:1 based on age and propensity score, and the other using the overall cohort with adjustment for propensity score--use of RAS antagonists were associated with significantly lower all-cause mortality, with a hazard ratio of 0.80 to 0.90. They also show a dose-response association, with a greater reduction in mortality with target doses of RAS antagonists than with lower doses.

They say that in contrast to randomized trials involving patients with heart failure with preserved ejection fraction, the patients in their registry more accurately reflect those treated in routine clinical practice.

Lund et al acknowledge that three previous randomized trials (CHARM-Preserved, PEP-CHF, and I-PRESERVE) have not shown significant benefits of RAS inhibitors in heart failure with preserved ejection fraction, but they say there were some signals toward benefit and that the trials may have suffered from selection bias and underpowering. In addition, their observational study included more severe heart-failure patients and a more unselected and representative population.

They conclude: "Our results together with the signal toward benefit in randomized trials suggest that RAS antagonists may be beneficial in patients with heart failure and preserved ejection fractions, but this should be confirmed in an appropriately powered randomized trial."

In his editorial, Fang points out that preserved ejection fraction in this study was defined by a left ventricular ejection fraction (LVEF) of 40% or more, which he says actually represents significant systolic dysfunction, and that the benefits appeared to be attenuated among patients with LVEF of more than 50%. But Lund et al argue: "Even though ejection fractions of 40% to 49% may not be considered normal, the benefit in this group has not previously been demonstrated."

The Perils of Confounding

Fang gives the usual caution on interpretation of observational data: "At the heart of the matter is whether any set of sophisticated statistical techniques can ever fully account for confounding in an observational study. Variability among healthcare practitioners, time-dependent changes in health status and medical therapy, and differences in socioeconomic status are all important sources of confounding that are not fully accounted for by measured covariates and not only influence the decision to use specific drugs but are known to independently be related to outcomes." He adds that "examination of large databases for signals of beneficial treatment effects in disease subgroups should not be considered definitive."

Asking how the findings from these two studies can be reconciled with those from randomized clinical-trial findings, Fang says the two types of trial provide complementary evidence, but that randomized trials remain the gold standard for assessing the true effect of an intervention.

In this case, he concludes: "If all of the evidence is carefully considered in its totality, it would be sound to conclude that (1) renin-angiotensin system antagonists are reasonable agents to control hypertension in heart failure with preserved ejection fraction, and (2) aldosterone antagonists are effective drugs in heart failure with reduced ejection fraction but should be used carefully and selectively. Although clinical trials should remain the gold standard for testing hypotheses, observational studies bridge the gap from the scientific rigor of clinical trials to real-world experience. Clinical trials are a reminder of the rigor of medicine as a science; observational studies are a reminder that medicine is still an art."