And now for a little tooth fairy science on meditation…

Back in the day, Deepak Chopra used to be a frequent topic of this blog. He still pops up from time to time, such as when irony meters everywhere immediately self-destructed after Chopra criticized Donald Trump for being insufficiently evidence-based or when, after I wrote a post asking why medical conferences keep inviting Chopra to speak, Chopra was so displeased that he actually posted a video attacking me (and other skeptics who’ve criticized his pseudoscience). Unfortunately, Chopra truly is one of the most influential people in “integrative medicine” today. To be honest, I’ve never really been able to figure out why. For the last few decades, no one has been better at using and misusing the word “quantum” than Chopra: Quantum consciousness, quantum medicine, quantum everything. His woo has been so powerful and easy to predict, that I even gave it a name, lo, these many years ago: Choprawoo. It’s hard to believe it’s been over a decade since I originally coined that term. It still fits.

In any case, in the wake of our little kerfuffle a few weeks ago, Chopra has (sort of) been making nice to me on Twitter. I rather suspect that he thinks he can win me over. Or maybe he thinks he can entice me into a public debate (not gonna happen—I’m not Michael Shermer—although I am always up for dialogue of Chopra’s serious). However, in exchanges on Twitter, one of his acolytes posted a link to a study that Chopra’s followers seem quite smitten with. You might remember two years ago, when I was dismayed that actual scientists took Chopra’s nonsense seriously enough to actually collaborate with him. As I put it at the time: Deepak Chopra tries his hand at a clinical trial. Woo ensues. Well, this is the woo that ensues, only it’s dressed up in fancy state-of-the-art genomics and next generation sequencing. It turns out that the link I’ve been seeing to a study appears to be one of the first publications to come out of Chopra’s wooful trial. It was published in August by Epel et al in Translational Psychiatry and entitled Meditation and vacation effects have an impact on disease-associated molecular phenotypes. The list of institutions is, alas, fairly impressive: UCSF, Capella Biosciences, Mount Sinai School of Medicine, UCSD, Harvard. I guess it just goes to show how deeply entrenched quackademic medicine is.

Before I discuss this study, let’s step back a minute and recall what I wrote two years ago about Chopra’s study:

This is what we in the biz call a fishing expedition. There is no real, concrete, testable hypothesis here, other than that Chopra’s woo-packed Ayurveda program is a good thing.


He doesn’t say “test the hypothesis” or “see if this program results in X and Y.” He doesn’t know what he’s looking for before he does the experiment. He only knows that it’ll be good. Rather, he says things like “prove scientifically” that his program does all sorts of wonderful things that aren’t specified in concrete, measurable ways. Instead, he’s going to shotgun measure a whole boatload of markers and endpoints, including next generation whole genome sequencing. If adjustment for multiple comparisons is not undertaken, I can virtually guarantee that this study will be “positive” in that it will find a “statistically significant” difference in at least a few markers. Because a “whole program” is being tested, it will be impossible to tell if any apparently beneficial changes observed are due to exercise (yoga), meditation, or to the special diet that participants will be consuming while at the Chopra Center.

With that assessment in mind, which proved prescient (although it wasn’t exactly difficult to predict this), now let’s look at the study. Here’s the rationale:

Ancient practices such as yoga and meditation have long been thought to combat stress and promote longevity, although empirical evidence for effects on aging processes under highly controlled experimental conditions is lacking. Further, it is inherently difficult to assess effects of meditation apart from simple relaxation. Advances in the understanding of the biological bases of aging enable better assessment of acute effects of salutary interventions on biomarkers of aging. For example, impaired regulatory systems leading to systemic inflammation, and excessive stress responsivity, are related to biological aging and may partly underlie pathogenesis of cardiovascular and Alzheimer’s diseases (AD). More recent systems biology approaches have identified gene regulatory networks associated with a diversity of biological processes, including immune and stress responses, and objectively linked them with disease or salutary states.

Leading to:

Here we examined how exposure to a short-term intensive residential meditation retreat affected biomarkers of aging and more general regulatory networks defining a wide array of biological processes. A residential retreat provides intensive daily exposure in a controlled environment but has the added ‘vacation’ effect of taking people away from the demands of their daily lives, which alone might affect regulation of stress pathways. Therefore, it is critical to compare the effects of a meditation retreat with an active randomized control group. Because regular meditators may have differences in brain function and structure, as suggested by meta-analyses,28 and greater changes in GE than novices, after meditation, we also recruited a third comparison group of experienced meditators. Our design allowed us to study the effects of meditation independent of the vacation effect, as well as to compare the effects of acute intensive meditation in regular meditators versus those newly trained in meditation.

In brief, this study examined women between 30 and 60 undertaking a one week retreat offered regularly by the Chopra Center for Wellbeing at a vacation resort (OMNI La Costa Resort and Spa). Women from this group who were “non-meditators” were randomized to the vacation-only group (consisting of the resort only) or to the retreat (referred to as the novice meditation arm). These women were then compared to a group of experienced meditators. The overall idea was to try to separate the “vacation effect” (effects due solely to relaxation and being away from the cares of everyday life) from any specific effect of meditation while analyzing a boatload of biomarkers in the form of blood levels of inflammatory cytokines and the many thousands of gene expression levels measured by whole genome sequencing. Here’s the CONSORT diagram (a standardized manner of describing randomization in clinical trials):

Consort diagram

The interventions were described thusly in the supplemental information:

  • Meditation Groups program. The Seduction of Spirit Retreat, led by Dr. Deepak Chopra and colleagues, at the Chopra Center, Carlsbad, CA, over the past decade, is a meditation retreat attended by several hundred participants per event. Its goal is to promote an intensive period of learning and psychological change. The retreat provides training in meditation (primordial sound meditation, which is similar to mantra meditation), foundations of yoga, and sutra, with the aim of promoting inner calm, expanded awareness of one’s body, breath, and self, and life-transforming skills. The retreat group had lectures and activities for the full day. Over the four days, this added up to 12 hours of meditation (4 times a day), up to 9 hours of yoga (2 times a day), and several lectures and, interactive self-reflection exercises each day.
  • Vacation (Control) Group program. The Vacation Group had an afternoon lecture on health behaviors each day for 1.5 hours and an optional activity in the morning, such as a leisurely walk. While dietary intake was not strictly monitored, both groups were served the same meals at the same dining center.

Subjects were assessed by various instruments designed to examine psychological measures of well-being at day 1, day 5, and ten months later. They also had blood drawn at day 5 for various molecular measurements. One massive flaw in the study that I noticed right away is one that I almost always find in studies like this. Can you guess what it is? That’s right! There was no blinding, or, as was described in the supplemental information section, “The intervention groups were not blinded, and site investigators and study personnel knew to which group participants had been randomized.” Sorry, Deepak. That’s a big no-no, for so many reasons. Not surprisingly, meditation was associated with an increased feeling of well being.

There’s another problem here. While the first two groups represent a fairly typical randomization scheme, the third group is is not comparable. After all, there might be a number of things associated with being an “experienced meditator,” such as differences in diet, exercise, lifestyle, and the like. Minimal effort was made to control for confounders; all that could be said was that the three groups were comparable in age, racial distribution, weight, and body mass index. that’s not enough.

For all three groups, besides blood drawn was subjected to telomerase activity assays and assays for inflammatory cytokines (substances secreted that promote inflammation). Telomerase, as you might recall, is an enzyme that adds on to telomeres, which shorten with each cell division. Telomere shortening is, of course, associated with aging and death. On the other hand, elevated telomerase activity is associated with cancer. RNA was also isolated from the blood of the subjects and gene expression analyzed by RNASeq, which is also known as whole transcriptome shotgun sequencing. Basically, RNASeq uses next-generation sequencing (NGS) to reveal the presence and quantity of every RNA sequence in a biological sample at a given moment in time. That’s tens of thousands of different RNA sequences.

As you might expect the results were mixed. For instance, for telomerase activity, there was no difference between baseline (day 1) and post (day 5) readings for the vacation group or the novice meditation group. Just the experienced meditator group showed any difference, and even then there was only a 23% increase. How significant this is biologically, who knows? In fact, this is a pattern I noticed. Either there was no statistically significant difference between baseline and post, or there was only a difference observed in the regular meditators. Many of these differences were marginally statistically significant.

If you want to get an idea of what this study is about, this passage should tell you all you need to know:

The vacation group showed a statistically significant increase in circulating TNF-α compared with the regular meditators, and a marginally significant increase when compared with novice meditators. Reasons for this increase could include those in the vacation group having an acute stress-related inflammation response to the blood draw, overexposure to sun or some other exposure that can lead to acute physical stress such as exercise. All participants ate a healthy ayurvedic diet, which is thought to be anti-inflammatory.

At this point, the investigators are just pulling stuff out of their nether regions. After all, the hypothesis (at least as far as I can tell given how vague is that vacation should decrease stress and that meditation should decrease it even further. Yet here we see that a major pro-inflammatory factor, TNF-α, was increased in the vacation group. OK, I get it. This is sort of in line with the main hypothesis, even though that hypothesis was never explicitly stated.

Now here’s the thing. This study really doesn’t show all that much of anything. Those of you out there with more expertise in next generation sequencing and systems biology can confirm this for me, although I warn you: It will take some reading and time to figure this out. Most informative for this purpose are Tables S7 and S8, where Gene Ontology Tree Maps are compared. What ends up being compared is not what one would expect, and the differences found tend to be in areas whose relevance I can’t help but question. Gene Ontology Tree Maps are a system for cataloging genes according to their biological functions. For instance, it is noted that one group of genes is suppressed only in regular meditators: immune response and positive regulation of chemokine production. Other gene expression modules suppressed in the post measurements compared to baseline include oxygen transport, erythrocyte (red blood cell) differentiation, and hydrogen peroxide metabolism. In Table S8, the comparisons become even more—shall we say?—targeted.

But if you really want to to know how weak the findings of this paper are, don’t listen to me. Listen to its authors:

Our study has several limitations. The changes in biomarkers and GE patterns appear salutary; however, no studies have tracked these markers to see how well they predict disease or longevity in a healthy sample. The sample size is relatively small and powered only to detect medium-to-large effects. Replication of these findings with larger controlled studies will be necessary. In addition, comparing regular meditators to non-meditators is somewhat problematic, given regular meditators likely differ on many lifestyle factors that make such comparisons complicated. For example, diet, exercise regimen and stress-reduction activities such as meditation likely differ between these groups and may not only lead to baseline differences in molecular and higher-order systems, but may prime such systems to respond differentially to resort and meditation interventions. In addition, the meditation arm of this study was not strictly sitting meditation, but also included yoga postures and self-reflection exercises and lectures. Further, whereas our design did attempt to control for the vacation effect, we did not randomize regular meditators to the vacation and meditation arms, making it difficult to interpret the changes observed in the regular meditators, given they were compared with a control group comprising non-meditators.

In fact, even taking the most generous interpretation, as the authors do, all they could find was what they refer to as a significant “vacation effect” that benefits all groups. They could only identify anything resembling a “meditation effect” within the regular meditator group. Of course, given how short term this study is (only a week) it will aways be appropriate to question the clinical relevance of any of the findings in this study.

Basically, this is a study that purports to find differences in gene expression that might or might not be clinically relevant, with the emphasis being on “probably not.” If you look at it critically, you’ll see that basically no significant difference was found between the vacation group and the novice meditators, and that most of the effects described appear to be in the experienced meditators, who aren’t really comparable to the other two groups. In other words, this is a fancy, expensive study that didn’t find much of interest but is being sold as evidence that Chopra’s brand of meditation modulates the immune system. It is what I like to refer to, inspired by Harriet Hall, as Tooth Fairy science, a term that generally refers to studies assuming a phenomenon not in evidence, such as the tooth fairy. As Harriet Hall likes to put it, you could design your study to determine if the Tooth Fairy leaves more money for a tooth left in a plastic baggie under the pillow than for a tooth wrapped in a piece of tissue (as we used to do in our family). Or you could look at the average amount of money left behind for the first baby tooth to fall out compared to the last tooth. Or perhaps you might attempt to correlate Tooth Fairy proceeds with the income of the toothless kid’s parents.

The point, of course, is that the Tooth Fairy doesn’t exist. None of this is to say that meditation might not have some health benefits. It’s just to say that this study is not good evidence that it does.