Is Cardiology Overly Enamored of New Unproven Therapies?

Cardiology Tops the List of Medical Reversals

John M. Mandrola, MD


July 12, 2019

Sometimes helping people is easy: antibiotics for bacterial infection, pacemakers for heart block, and stents for acute myocardial infarction. But medical practice is rarely that black and white. Most of what we recommend in the name of helping people is of uncertain benefit.[1]

A novel study published in the open-access and transparently peer-reviewed journal eLife attempted to quantify how often our efforts to help people prove to be of no benefit.[2] This look back at medical history deserves attention.

Vinay Prasad, MD,* an oncologist at Oregon Health & Science University, and his team have spent thousands of hours describing the problem of low-value care. In this paper they studied the presence of medical reversals, which are a subset of low-value care defined as an accepted practice that is found through randomized controlled trials (RCTs) to be of no benefit or even harmful.  
The eLife paper extends work that Prasad and colleagues published in 2013. For that study they scanned the New England Journal of Medicine(NEJM) for RCTs that overturned common practices and found 146 examples.[3]  Now the authors extend their search for reversals to three journals: NEJM, The Lancet, and JAMA.

They identified more than 3000 RCTs published between 2003 and 2017. They excluded studies of novel therapies (for which there could be no comparison) and trials that supported the conventional practice or yielded inconclusive results.

That left 396 RCTs with results that did not support a common medical practice (13%).

A past criticism of this work is that their definition of a medical reversal may differ from that used by content experts in that area. To address this, Prasad and colleagues searched for systematic reviews of any therapy they deemed subject to a medical reversal. The systematic review confirmed their call in 209 cases, slightly more than half.

Medical reversals covered many areas of medicine, with cardiology topping the list (n = 80, or 20%), followed by public health/preventive medicine (n = 48, or 12%) and critical care (n = 45, or 11%).

A key feature of this paper is the 400-page evidence table in the appendix, where they list all the reversals.

Here are four highlights from the reversals identified in cardiology.

  • The bioresorbable vascular scaffold (BVS) or dissolving stent: The BVS had extreme levels of plausibility, and even US Food and Drug Administration (FDA) approval, yet RCTs proved it to be not only inferior to metallic drug-eluting stents but also harmful.[4,5,6] I add this reversal because FDA approval does not inoculate a practice from being reversed.

  • Intra-aortic balloon pump (IABP) for cardiogenic shock and high-risk percutaneous coronary intervention (PCI). IABP use for these two indications was once common, but now a meta-analysis (for cardiogenic shock) and clinical trial (for high-risk PCI) have reversed this practice.[7,8] I highlight the IABP reversal because it is connected to the controversy over newer types of mechanical support, such as Abiomed’s Impella, an expensive and popular FDA-approved device. The problem: the Impella vs IABP trial during high-risk PCI was stopped for futility,[9] and there are no trials of Impella in cardiogenic shock.

  • Automated chest compression during resuscitation: I recently saw one of these devices in action in the emergency department. The violence of it struck me. “Do those things really work?” I asked a paramedic. “Oh, yes,” he said without wavering. He was wrong. A large RCT found that the automated devices were associated with worse neurologic outcomes compared with manual compression.[10] A systematic review confirmed the negative RCT.[11] I add this reversal because our city stocks the ineffective devices in its ambulances—years after the publication of the unsupporting data.

  • Ablation beyond pulmonary vein isolation (PVI) for atrial fibrillation (AF) ablation: In the STAR AF 2 trial, extra ablation outside the basic PVI lesion set did not reduce the rate of recurrent AF in patients with persistent AF.[12] I was among those duped by key opinion leaders who promoted the idea that advanced forms of AF require more ablation. Before de-adopting the practice, I spent years doing repeat procedures for arrhythmias created by the extra lesions. If only I had been a medical conservative[13] earlier in my career.

Space limits my expanding on the hundreds of reversals outside of cardiology, but I came away stunned by the vast scope of these. Failures of preventive measures (e.g., transmission of resistant bacteria in the hospital, peanut allergy in children, and report cards for cardiac care) and the numerous well-meaning screening programs caught my attention.


The main limitation of this paper comes in the judgment of calling a practice reversed.

For instance, the authors cite a negative RCT of palliative care–led meetings for families of patients in the intensive care unit (ICU).[14] Trial authors concluded that these meetings were no better than usual care for relieving anxiety or depression symptoms. This reversed a clinical policy statement[15] recommending that palliative care specialists offer support for families of patients in the ICU. But a palliative care physician who participated in the trial objected to the conclusion on methodologic grounds.

Another reversal I would push back on is that intensive lifestyle intervention in patients with type 2 diabetes does not reduce cardiac events. This call came from the Look AHEAD trial, which found that an intensive lifestyle intervention did not reduce cardiac outcomes in overweight/obese adults with type 2 diabetes.[16] The problem was that the average difference in weight between the study groups was a mere 4 kg (about 9 lb). That’s not enough. Scottish authors report that a more intense weight loss program led to remission of diabetes in a third of people with established disease.[17]

Note that I pushed back on two topics I feel strongly about: palliative care and lifestyle measures. Other clinicians might push back on reversals in their field. This underscores the notion that while many reversals are obvious and agreed upon (treating premature ventricular contractions with antiarrhythmic drugs after myocardial infarction), many have gray areas.

Then there is the judgment of whether a reversed practice was commonplace. The authors cite a reversal in the use of fibrinogen concentrate to reduce intraoperative bleeding during heart surgery.[18] My surgeons say they have not routinely used this practice. Similarly, the authors list multiple reversals in the use of the calcium-sensitizing drug levosimendan for improving outcomes after heart surgery, but the drug is not used in the United States.


Medical knowledge changes over time; thus, some degree of medical reversals is expected.

Cardiology has a large number of available therapies. We often put them to the test, so it’s not surprising that we had the most reversals.

Some might argue that a 13% rate of medical reversals is acceptable, even normal. I would agree with that if we were slower to embrace unproven therapies and faster at de-adopting ineffective therapies. We are not. The issue is not the percentage of reversals, it’s the scope of low-value care that existed before and persists after a reversal.

Measuring only 26 low-value services for Medicare-aged patients, Schwartz and colleagues estimated costs in the billions of dollars over only 1 year.[19] That analysis is 10 years old, and, given the rising costs of care and the hundreds of reversals, the money we waste now in low-value care is shameful.

An obvious takeaway from this research is that people who advise patients must become better judges of evidence. Our urge to help people makes it easy to be fooled. Most of the reversed practices became established on weak evidentiary grounds—a flawed RCT, single-center studies, observational data, or, the most fragile of them all, expert opinion.

Medical reversals also reiterate the clinician’s need to fight against overconfidence. While we are certain about a few practices, uncertainty pervades most medical decisions. As a thought experiment, ask yourself: Did my patient get better because of the treatment, or despite it? I used to think it was the former; I now think the latter happens more than we realize.

Finally, this study of medical history supports the tenets of medical conservatism.[13] Human nature holds that in the face of uncertainty, patients and doctors have a bias toward action—do more, not less, like I did in embracing extra ablation for persistent AF.

To observe, to do less, or to discuss uncertainty can seem scary. But this paper teaches us that if a practice does not have clear evidence of benefit, doing less may lower the odds that the next generation of clinicians will consider us fools.

*Additional Disclosure: John Mandrola has coauthored papers with Vinay Prasad. Prasad is also a Medscape contributor.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: