Curmudgeon’s Corner

Can We Hurry Up the Truth in the COVID-Era?

Jacob Schor, ND

There’s a new trend in medical research publishing that we are seeing more often in the last weeks. Scientific studies are being released as pre-prints before they have undergone formal peer review or revision with the ‘understanding’ that they will eventually be published the normal way. Basically, what we see is the first draft.

The advantage to this is that urgent information is made available to other researchers, medical practitioners and the public sooner rather than later. COVID-19 has brought a layer of urgency to our world, a sense of now or never, and we want the information as soon as possible without the inherent delays built into the reviewing process.

The disadvantage is a kind of ‘caveat emptor’. We are left wondering if the findings are actually true or not. One could say that this is always the case, that just because information is published in a peer reviewed journal is no guarantee; but now this is even more the case.

Two separate preprint papers have been sitting on my desktop. The first to arrive was a preprint of William Grant’s “Vitamin D Supplementation Could Prevent and Treat Influenza, Coronavirus, and Pneumonia Infections.” This was received by the website PrePrint on March 12, posted March 15, and as of April 2 is now published in the journal Nutrients.¹

Most of us are familiar with William Grant or at least his work. He’s been one of the driving forces behind the, what do shall we call it? the Vitamin D Revolution? or the Vitamin D Fad? You can review his ideas on the Vitamin D Society’s (https://www.vitamindsociety.org) or on SunArc.org. Most of us are quite familiar with the message. This paper clearly presents their side of the topic. Nothing wrong with having a bias, well, except what gets left out. Reading this paper one could easily believe that all of the published vitamin D research supports Grant’s view of reality. They neglect to mention the multiple meta-analyses that sum up past RCTs trials that testing vitamin D to prevent disease as “futile.”^{2 3 4 5}

Grant recaps Cannell’s original 2006 paper that suggested vitamin D would protect against seasonal influenza.⁶ Cannell’s thesis relied heavily on Edgar Hope-Simpson’s book, The Transmission of Epidemic Influenza,⁷ and this inspired me to pull it from my shelf to page through. I recall how influential this book was: much of it is underlined in yellow highlighter, I must have hoped to commit those lines to memory.

The actual clinical trial described in the Grant paper is an analysis of online questionnaires completed by individuals who purchased vitamin D testing kits through GrassRootsHealth.net, another group whose mission is to promote vitamin D.

Test purchasers whose tested vitamin D levels were 60 ng/ml or above had lower risk of reporting that they had the flu in the prior six months compared to those whose test showed vitamin D levels of 20 ng/ml or less; reported infection rates dropped from 12% to 8% (N = 12,605). The authors reported this as a 43% change in risk, which makes it sound like a lot more. This was an open trial. It does not tell us whether low vitamin D status increases risk of disease or whether vitamin D levels are reactive to other factors that influence disease risk. Serum iron levels are an example of a reactive nutrient that drops rapidly with infection, cancer, or other inflammatory conditions.

Grant et al then argue that vitamin D will be protective against COVID-19 because the epidemic in China started in the fall when vitamin D levels drop and that prior epidemics caused by other coronaviruses show seasonal variations. The best argument in support of Grant’s theory will be whether COVID-19 infections also exhibit a seasonal variability. Perhaps by the time you read this, the data will tell us if infections rates decrease in the summer or not, the way influenza does.

The second preprint paper on my desktop is Wu et al’s “Exposure to air pollution and COVID-19 mortality in the United States” that was released in early April 2020.⁸ This team effort comes to us from the research lab of Francesca Dominici of the T.H. Chan School of Public Health at Harvard University. This group has produced a series of research papers in recent years focused on fine particulate air pollution and have already earned our respect in this specialty.

The Chan researchers collected air pollution data for 3,000 counties in the United States right up to April 04, 2020, the day prior to the paper’s release, and then compared county level COVID-19 deaths against county level long-term average PM2.5 exposure. Adjustments were made by population size, hospital beds, number of individuals tested, weather, and socioeconomic and behavioral variables, including but not limited to obesity and smoking. Their compiled data covered 98% of the US population.

Wu’s team calculated that an increase of 1 μg/m³ in PM2.5 is associated with a 15% increase in the COVID-19 death rate, [95% CI:5%, 25%]. Results are statistically significant and robust to secondary and sensitivity analyses. 

In our lifetimes there has not been a public health emergency that matches the COVID-19 pandemic. Both the suddenness and global scope of this disease has created great urgency to identify modifiable factors that will affect disease spread and outcomes. Thus, we are in a hurry to find ways to shift disease morbidity and are willing to look outside the normal published literature to find answers or at least clues to what might help.

Many of the risk factors already known to increase risk of mortality from COVID-19 infection are conditions that are known to be worsened by PM2.5 exposure and that are associated with low vitamin D status. These include diabetes, hypertension, heart disease and others.

The Global Burden of Disease Study identified air pollution as a major risk factor for total and cardiovascular disease mortality and, in 2015, suggested it was responsible for 5.5 million deaths worldwide a year.^9,10 These various groups promoting vitamin D have probably produced a similar calculation although recent studies have questioned earlier predictions.

The findings of the air pollution study are relevant for several reasons. The first is that we can reduce the risk of severe disease in high-risk populations by lowering the burden on their health due to air pollution. While ambient pollution levels are beyond an individual’s immediate control, indoor air quality can be rapidly improved by using a home air filter. The second reason this information is relevant is that any move that will lower outdoor air pollution may lower mortality rates during the pandemic and should be encouraged. Conversely, any increase in outdoor air pollution may increase mortality rates. This should inform our assessment of the March 26, 2020 announcement from the EPA that they have instituted a “… sweeping relaxation of environmental rules in response to the coronavirus pandemic, allowing power plants, factories and other facilities to determine for themselves if they are able to meet legal requirements on reporting air and water pollution.”

The vitamin D study is harder for me to interpret as, at this point, the proposition that vitamin D will impact COVID-19 feels more tenuous. The actual study as described sounds weak and the authors could not be more biased. Still, the risk associated with supplemental vitamin D appears small. The change in infection risk, if it even applies to COVID-19, also seems small. The impact that our nationwide efforts at social distancing will have on vitamin D status is yet to be determined. Will levels increase as people work from home or will they drop as people eschew outdoor activities? At this point we can only guess.

Let’s try to unwrap these numbers more carefully. Harvard’s air pollution researchers tell us that an increase of only 1 g/m3 in long-term average PM2.5 is associated with a statistically significant increase of 15% in the COVID-19 death rate. In New York City if an intervention were to lower the long-term average PM2.5 exposure by only 1 g/m3, there would have been 248 fewer deaths up to April 4, 2020, when the total was 1905—or 861 fewer as of today, April 11, when the total is up to 5,742. As I wrote this, the PM2.5 level in NYC is 28 g/m3. It has fluctuated from 10 to 58 g/m3 over the past 24 hours. In China, pollution levels dropped 25% during the lockdown, so conceivably NYC could see a similar drop of 2 to 14 gm/ m3. We might be talking about rather serious shifts in mortality risk. There is ample room for additional improvement.

In a previous study published in 2017, these same authors examined data from 61 million Americans Medicare Beneficiaries and found that an increase of 1 mg/m3 long-term PM2.5 exposure was associated with a 0.73% increase in the rate of all-cause mortality.¹¹ Therefore, a small increase in long-term exposure to PM2.5 leads to a large increase in COVID-19 death rate of a magnitude that is 20 times the one estimated for all-cause mortality.

Air pollution levels were also associated with increased risk of death during the Severe Acute Respiratory Syndrome (SARS) outbreak in 2003, which was also caused by a type of coronavirus.¹²

If Grant’s vitamin D predictions are accurate and we were to distribute massive doses of vitamin D to large segments of the US population, we might see a 4% reduction in risk of flu infection. Trying to infer what this would do to COVID-19 infection rates and, more importantly, mortality rates feels like too big a stretch given the weakness of these data. Still it can’t hurt to encourage supplementation and, perhaps more important, getting some sun exposure.

That’s all assuming that these draft papers do not contain major errors in their calculations. I can’t say that for sure. Wu’s calculations are far beyond my mathematical skills to follow. There were obvious typographical errors in the released draft that I read, but the reputation of the authors makes me want to assume that their main findings are unbiased and accurate.

The preliminary status of the air pollution paper hasn’t prevented the conclusions from being picked up and broadcast worldwide both on the news and social media. Does that make them fake news? More like premature news. Yet where do we draw the line? Probably the same way we always do when dealing with uncertainty of medical efficacy. We fall back on Auguste Chomel’s now universal¹³ (not Hippocrates) proscription of primo non nocere; we look for potential harm and balance that against possible benefit. Breathing clean air or basking in sunlight seem to pose little danger, while both might offer possible benefit.

____________________________________________________

SIDEBAR:

What Is a Preprint? And Why Do We Care?

These two articles mentioned were posted on bioRxiv, pronounced “bio archive.” This site was first launched in 2013 and is overseen by Cold Spring Harbor Laboratory, an old and highly respected research institution that was founded in 1890. BioRxiv is modeled after arXiv, (pronounced Archive) a preprint server focused on the physical sciences that was begun in 1991. The hope that inspired creation of these websites was that rapid dissemination of biological research would help other researchers. A third preprint site, medRxiv (as in Medical Archive) was launched in July 2019 to focus on health sciences.

With the COVID-19 pandemic in full force, utilization of these preprint services has grown with the urgency to share new findings on the disease. Downloads from medRxiv increased 100-fold between December 2019 and April 2020. About 70% of papers posted on bioRxiv and arXiv have ended up being eventually being published. MedRxiv is too new to have similar statistics on yet.

There is a low bar to get a paper posted. To have a paper posted on bioRxiv or arXiv, all that is required is for all of the authors to consent and declare that their study had ethical approval and participant consent and state any conflicts of interest and if the study was a clinical trial.

Once uploaded the paper is screened to be sure it wasn’t plagiarized. BioRxiv prohibits health claims and medRxiv won’t accept case reports or trials with small sample sizes. In theory, both servers refuse papers that may encourage dangerous behavior. That’s pretty much it. In the past this was enough.

Problems have started to crop up though. News editors and writers are eager to write about new science and will quote preprint papers as if they were actually published studies. This happened in early April 2020. The New York Times mentioned the study on fine particulates being associated with COVID-19 mortality rates in an article describing how the Trump administration halted plans by the EPA to tighten pollution standards, “Harvard University this month also published research showing that the coronavirus causes a higher death toll among patients in parts of the country with increased levels of fine particulate pollution.”¹⁴

Technically this study was not published. It was only posted on bioRxiv. There is an important difference.

We see repeated examples of individuals embracing and promoting the information read in a preprint either from ignorance or malicious intent. Preprint papers have been posted online and used to fuel conspiracy theories.

“Science is a conversation,” said Dr. Ivan Oransky, one of the founders of Retraction Watch, a blog that reports on retractions of scientific papers. “Unfortunately, people in times of crisis forget that science is a proposition and a conversation and an argument. I know everybody’s desperate for absolute truth, but any scientist will say that’s not what we’re dealing with.”¹⁵

While we are eager for information related to COVID-19, we need to be extremely cautious if we are reading preprints rather than published studies.

This has become increasingly important as published science for many of us has become a defensive line against ‘fake news’, that is the spread of falsified information that is amplified by social networks online. For those of you who still think this isn’t a significant concern I would suggest reading Soroush Vosoughi’s study on the spread of fake news published in Science two years ago.¹⁶ Or at a minimum read the excellent summary of the study by Robinson Meyer published in The Atlantic.¹⁷