Survey results presented at the American College of Cardiology 2019 Annual Scientific Session led to a declaration that went well beyond the evidence - they said electronic cigarettes can lead to a dramatic increase in the odds of having a heart attack, coronary artery disease, and even depression, and used that to state, "These data are a real wake-up call and should prompt more action and awareness about the dangers of e-cigarettes." 

Except "these data" did not show that, and they were forced to admit it.

The results were determined by collating results from the 2014, 2016, and 2017 National Health Interview Survey, which is funded by the Centers for Disease Control and Prevention. Some of the results make obvious sense and are in line with the consensus, like that smokers were 78 percent more likely to have had a stroke and 165 percent more likely to have had a heart attack. That won't get any attention, no one outside Altria marketing is on Team Cigarette in 2019, everyone knows smoking kills you.

What will get attention is one of the CDC's pet causes - their war on vaping. Along with promoting "prediabetes" - we're the only country in the world who touts this - and that opioid issues have been created by doctors, pain patients, and pharmaceutical companies rather than recreational users, the CDC has been going after smoking cessation (quitting) and harm reduction (fewer cigarettes) tools like e-cigarettes. Their rationale in the past has been as suspect as their claims about prediabetes; if a rebellious teen used an e-cigarette in the last 30 days, they were considered a chronic user.(1)

Using survey results, the scholars at the ACC meeting declared that e-cigarette users were 56 percent more likely to have a heart attack and 30 percent more likely to suffer a stroke than non-users. Bizarrely, they were twice as likely to recall having depression or something on the spectrum of emotional problems.

The weakness is that the results were all self-reported and subject to recall bias, and the data only allowed the scholars to control for age, sex, self-reported BMI, high cholesterol, high blood pressure(2), and...smoking. When the data are just limited to the small pool of vaping users who are not also current or former smokers, there was basically nothing to talk about.

A third of the vapers also still smoke

Vaping is used in both smoking cessation and harm reduction so concluding that vaping causes an increased risk of heart issues without eliminating smokers is just wrong. Had the authors done that, they had no results with any confidence.

Vaping did not even begin to get popular until 2014, it was really only available starting in 2007. It was still such a minor market that in 2015 the CDC did not ask about it on the survey at all (which is why the years 2014 and 2016 were used though not 2015) but somehow, in defiance of everything we know about biology and toxicology, the authors set out to claim it already caused spikes in stroke.

To believe that, we have to ignore settled science. Damage from smoking builds over time, you don't smoke a cigarette and get lung cancer the next day, and because vaping lacks the 200 toxic carcinogens in cigarette smoke it would take even longer for diluted nicotine vapor to come anywhere close to smoking's harm, if it can happen at all. It did not happen in three years.

In the US, e-cigarette users are a tiny population and while the majority of those are former smokers, about 20 percent still smoke. Had the authors factored out confounders, such as being a former smoker or a current one, they really only had a few hundred people - 500, senior author Mohinder Vindhyal, MD, of the University of Kansas School of Medicine admitted to Crystal Phend at MedPage Today. "With those 500, we could not get any meaningful data or outcomes or associations."

So the authors touted statistical significance of P<0.0001 without choosing the correct data. 

Statistical significance is misused and this is a great example

In Nature, I am one of the signatories of an article called Scientists rise up against statistical significance, and it is hopefully the beginning of a movement that's been needed for 30 years. Ever since epidemiology had its big win, smoking, there has been a rush into statistics to try and duplicate that fame. So now we have statistical claims about diet soda, Manwich, bacon, hot tea, weedkillers, you name it and some epidemiologist has linked it to cancer, even though actual human data do not show it. And in some cases it is biologically impossible. 

An ideologically neutral examination of these survey results would be tempered with the reality that former smokers who use vaping tools for cessation or harm reduction, or current smokers that don't yet have health issues, could still develop them with age. Health issues may even be why they quit smoking. At the same ACC conference it was noted that the rate of adults under 40 having a heart attack has risen 2% every year for the last 10 years. Genetic flukes or bad luck aside, the reason for that is obesity, the number two lifestyle killer. Combined with the damage from smoking, the number one killer, they are using a curve in popularity to claim a curve in health effects.

But it's a spurious correlation when done that way. 

Even though they can't legitimately claim harm they still attack the soft underbelly of public trust by claiming vaping is not "safe", as in paragraph one. Nothing is safe but conservative studies show vaping is 95 percent safer than cigarettes and that is fantastic.  Yet somehow in America the government fad is against vaping whereas in the UK their government-run NHS recommends it for smokers. 

Fads are not how health policy should be made, but too often journalists help it to happen. Three years ago we were told sitting at a desk was as bad as smoking  - using similar survey results and statistical correlation. It was a terrific subsidy for the standing desk market, brought to them by government, sensationalist journalism and academia, but it didn't improve public health one bit. A war on vaping, which gets credit for the decline in smoking, is a mistake unless the data become clear it is truly harmful.

The US Food and Drug Administration has been a lot more neutral on smoking cessation. Under Commissioner Gottlieb they got rid of the Obama administration's arbitrary date for product registration, which would have created 10 million casual criminals and led to lawsuits, while hammering suspect companies who are targeting kids. With his departure, the anti-smoking community is hopeful he'll be replaced by someone more like him and less like CDC Director Robert Redfield, a great virologist who has punted smoking cessation issues to Obama holdover Anne Schuchat, who claims any amount of nicotine is as bad as cigarettes.


(1) CDC is also plagued by willingness to turn over their positions to medical societies which are often biased toward their members. With pre-diabetes, ADA runs their panels. We recommend prostate-specific antigen (PSA) screening for healthy men aged 45 and older, something no other country recommends at all, because 75 percent of the guideline panel for prostate cancer is in the same urology organization. With smoking cessation, anti-smoking groups which have seen their Master Settlement funding expire need a new target, so they have become anti-tobacco. And then pharmaceutical companies that make competitors to vaping, like gums and patches, are happy to help behind the scenes.

(2) High cholesterol and high blood pressure are not diseases, they are risk factors linked to diseases. So these authors used a risk factor for a disease to create another risk factor for the disease. None of this makes any sense to people from Missouri, Kentucky, or places where scientific common sense is used.