COMMENTARY

Jan 21, 2022 This Week in Cardiology Podcast

John M. Mandrola, MD

Disclosures

January 21, 2022

Please note that the text below is not a full transcript and has not been copyedited. For more insight and commentary on these stories, subscribe to the This Week in Cardiology podcast on Apple Podcasts, Spotify, or your preferred podcast provider. This podcast is intended for healthcare professionals only.

In This Week’s Podcast

For the week ending January 21, 2022, John Mandrola, MD comments on the following news and features stories.

Olive Oil

The top cardiology journal has published a study on olive oil. It’s the most read news story on Theheart.org | Medscape Cardiology. It’s been covered by more than 100 news outlets and Tweeted more than 200 times.

Before I tell you about it, I want to give a disclosure: I am half Italian and was raised in an Italian-dominated family but I don’t really like olives. Now to the study. A research team, led by the Harvard TH Chan School of Public Health, used the longitudinal databases of the Nurses’ Health Study and Health Professionals Follow-up Study to evaluate the association of olive oil intake with total and cause-specific death rates.

The format of these sorts of studies is familiar. They assess olive oil intake with food questionnaires; they break down olive oil intake into four categories of none, a little, a little more, and a lot of olive oil; they assess deaths; they calculate hazard ratios based on olive oil consumption; they make statistical adjustments. Since this is the lead article in the Journal of the American College of Cardiology (JACC) and since it’s received so much attention, you can probably guess the results.

The multivariable-adjusted pooled hazard ratio (HR) for all-cause mortality among participants who had the highest consumption of olive oil (> 0.5 tablespoon/day or >7 gms/day) was 0.81 (95% confidence interval [CI]: 0.78-0.84) compared with those who never or rarely consumed olive oil. That’s a 19% reduction in death!

Higher olive oil intake was also associated with:

  • 19% lower risk of cardiovascular disease mortality;

  • 17% lower risk of cancer mortality;

  • 29% lower risk of neurodegenerative disease mortality;

  • 18% lower risk of respiratory disease mortality.

In other words, olive oil shreds human diseases and proves to be a modern-day fountain of youth. Now, before we talk about the elephant in the room, bias, let’s first point out that olive oil has a number of potential healthy attributes: it has anti-inflammatory, anti-atherogenic properties; it may improve lipid profiles, reduce blood pressure, and lessen endothelial dysfunction. I have no doubt olives are better for us than Fritos.

But, my friends, if you believe that a self-reported delta of 0.5 tablespoons per day of olive oil leads to this degree of death, cancer, and dementia reduction, I have a left atrial appendage plug to tell you about.

Please do read the editorial from Dr. Susanna Larsson from Uppsala Sweden. She gently points out the problem of residual confounding, which is that people who self-report olive oil intake are healthier than those who don’t. To me, the fact that three of the most common causes of human death and disability were crushed by olive oil, strongly suggests correlation not causation.

You also have the issue of self-reporting dietary intake, which was done every 2 years. I mean, if I was asked what my olive oil, or canola oil, or butter intake was 2 years ago, it would be a total guess. I could tell you my pizza intake, but not much else. (My family eats pizza every Friday night.)

The authors use complicated statistics to try to adjust for confounding. The statistics are probably top notch, but you can only adjust for the data you have on the questionnaires and spread sheets, which is barely the tip of the iceberg when you think of all the causal factors that go into getting heart disease, dementia, or cancer over two decades.

In sum, by all means if you like olive oil, use it. But I don’t think you should eat it to lessen human disease. To me, the benefit of the Mediterranean Diet stems mostly from the fact that people who adhere to it eat a lot less processed Popchips.

Final question: While publication of such a study brings attention, clicks, likely citations, and a higher Impact factor, does it increase or decrease the scientific standing of a journal as a judge of clinical science?

Breastfeeding and CV Risk

Speaking of observational studies at risk of confounding, the Journal of the American Heart Association (JAHA) published a meta-analysis of eight studies with over a million women, looking at the association of breastfeeding and future cardiovascular (CV) outcomes of the moms. The moms who breastfed is one group, the moms who did not is the other group.

  • When they combined the eight studies, breastfeeding was associated with an 11% lower risk of maternal CV disease; the HR was 0.89 with a CI ranging from 0.83-0.95.

  • The risk reduction was 14% for coronary artery disease, and 12% for stroke; both were statistically significant.

If you take these at face value, you can use this paper to promote breastfeeding. You could say there is biologic plausibility for maternal benefits. And boom, we have major statistically significant reductions in a study with a million(!) subjects. But....

There are two major parts of interpreting a meta-analysis that this study highlights as learning opportunities. One is that when you do a meta-analysis, you apply something called the GRADE or Grading of Recommendations Assessment, Development, and Evaluation tool to assess the quality of the evidence of the included studies.

Thus, part of a meta-analysis is understanding the quality of the studies being combined. The authors did this and report that the included studies ranged from very low to moderate quality of evidence. That makes sense, because observational studies comparing women who breastfed with those who did not are going to be at risk of bias because women who breastfeed almost certainly differ in many ways from those who don’t breastfeed. Of course, authors try to adjust, but as I said earlier, adjustment can only be done for the covariates that are on the spreadsheet.

Another important part of interpreting meta-analyses in this study is heterogeneity. The idea is that when you combine studies, your confidence in the results turns a lot on how similar they are.

Heterogeneity is measured with something called an I-squared (I2), and you want it to be low. In this study, the I2 was nearly 80%. Meaning the authors were combining studies with results all over the map. A nice review from Italian authors discusses the pitfalls of meta-analyses. They write: Actually, when high heterogeneity is evident, individual data should be not pooled and definitive conclusions should be drawn when more studies become available.

I am no expert on breastfeeding. There are many great reasons for moms to breastfeed, and one of them may be better maternal CV health. But this data cannot and should not be used to promote breastfeeding. The biggest benefit of this data, I think, is in highlighting the limitations of observational data.

Informed Consent in Patients with Stroke

The journal Neurology has published new guidance on informed consent in patients with acute stroke. Journalist Sue Hughes has excellent coverage. The issue is important because patients who present with stroke may, by virtue of the brain ischemia, not have decisional capabilities. And of course, there is the need to work quickly with stroke.

The problem is that acute stroke therapy remains controversial, in my opinion. Sue’s news article includes this quote: He gives a classic example of standard of care as a patient presenting with an acute ischemic stroke within 3 hours of symptoms who meets the criteria for thrombolysis. "Even if they cannot speak and cannot consent that would be a classic example to proceed with thrombolysis as the indication is clear-cut...."

I strongly disagree with this statement on the grounds that a critical appraisal of the systemic lysis literature suggests that “classic” is hardly the right modifier for indication.

In 2018, I wrote a column making the case against IV thrombolysis in acute stroke. While many in the emergency medicine community agreed – in fact, the emergency medicine community has led the way in critical appraisal of this area — many in the neurology community strongly disagreed. I urge you to read both as a “neutral Martian.”

The issues are complex and beyond the scope of this podcast but include:

  • The original National Institute of Neurological Disorders and Stroke (NINDS) trial had baseline differences in stroke severity, and a patient-level analysis suggested that the small benefit seen with tissue plasminogen activator (tPA) was likely due to baseline stoke severity not tPA.

  • There have been many IV-tPA trials and most were null. Only two were positive (NINDS and the European Cooperative Acute Stroke Study).

  • The benefit, if there is one, is measured in a very subjective modified Rankin Score, which is not always reproducible.

  • IV-tPA clearly increases the risk of intracerebral hemorrhage (ICH).

  • IV-tPA leads to an early mortality signal.

Before you all get mad at me, I have backed off a bit from my 2018 stance, but at best, you can say the absolute risk reduction in subjective improvement with thrombolysis is small, and the upfront harm of ICH and early death is real. At worst, a commonly accepted therapy is of no net benefit. The take-home for me is that informed consent here is even harder when you look at the data closely. I guess you could say that about a lot of therapies. A final caveat, the randomized controlled trial-level data for endovascular therapy, when patients are eligible, is far more robust than for IV thrombolysis. Here, I think the informed consent process is a bit easier. This post makes the podcast because IV-thrombolytic therapy is one of the most intriguing areas of critical appraisal that I know of and because informed consent is so darn important.

Controversy in the Coronary Revascularization Guidelines

In December, the ACC, AHA, and the Society for Cardiovascular Angiography and Interventions (SCAI) published a new guideline for coronary revascularization. The guideline sparked serious controversy among surgical societies. First, the American Association for Thoracic Surgery (AATS) and Society of Thoracic Surgeons (STS) withdrew their support. More recently the Latin American Association of Cardiac and Endovascular Surgery (LACES) has demanded "urgent reconsideration."

The point of contention: the guideline authors downgraded the recommendation for coronary artery bypass graft (CABG) in patients with multivessel disease to reduce overall mortality from a Class 1 (the highest level) to IIb (the lowest level) recommendation. The surgical groups also find problematic the fact both CABG and percutaneous coronary interventions (PCI) garner the same IIb rating.

The four surgical arguments are instructive:

  • Surgeons point to the famous Yusuf metanalysis of CABG vs medical therapy from the 1994. This established surgery as superior to medical therapy in patients with multivessel disease.

  • Surgeons also cite SYNTAX, EXCEL, NOBLE, and FAME 3, which all point to the notion that CABG is superior to PCI for CV outcomes.

  • The 10-year follow-up of the MASS II RCT, which compared PCI, medical therapy, and CABG, =found a lower incidence of cardiac mortality (as part of its secondary outcomes) following CABG compared with optimal medical therapy and PCI.

  • Finally, the surgeons argue against the use of the ISCHEMIA trial to downgrade CABG. Recall that ISCHEMIA studied early invasive to delayed invasive angiography and revascularization in patients with serious ischemia. ISCHEMIA found no reduction in CV outcomes for the invasive approach.

I think this fourth point is the most important. What I think the guideline authors are saying is that ISCHEMIA included patients with serious CAD; and revascularization, which included CABG, had no effect on hard outcomes. So we should downgrade the indication to do CABG in patients with stable CAD and multivessel disease. Part of me agrees with them—the part that feels that stable CAD is really a stable condition.

Bill Boden, the principal investigator (PI) of COURAGE, frequently reminds me that docs think patients with multivessel disease are going to fall over dead if not “fixed.” But recall that in ISCHEMIA, there only 66 sudden deaths among nearly 5200 patients. The incidence was only 1.3%. Boden and colleagues, including the PI of ORBITA, Rasha Al-Lamee, have a great editorial on optimal medical therapy in the Lancet.

Medical therapy is pretty darn good these days. Sure, Yusuf found that CABG beat medical therapy in 1994, but medical therapy is a heck of lot better these days. But it is simply not correct to cite ISCHEMIA as a reason to downgrade CABG to the lowest recommendation based on early revascularization not improving outcomes in ISCHEMIA. Here only approximately 24% of patients received CABG; 76% received PCI. It’s a huge leap of faith to believe that outcomes of a small subgroup of ISCHEMIA means surgery should be relegated to the lowest level of recommendation. As far as I know, there are no publications looking at the surgery subgroup, but perhaps there shouldn’t be because the statistical power would be low and surely patients referred for surgery are inherently sicker.

In sum, I agree with the surgery pushback. It seems more reasonable to put surgery at a higher level than PCI. But here’s the thing: I wonder how much effect these guidelines will have in the United States. Here, practice patterns are entrenched.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
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.

processing....