In 2007 I got an email from Preston Estep, a gerontologist and former Chief Scientific Officer of Longenity, Inc., offering me a place in an informal trial of the benefits of resveratrol that he was organizing. Recently I wrote him to find out what happened. Here’s his reply:
We got a few people to volunteer but not enough for an organized trial to be worth the effort, partly because initial reported benefits evaporated under scrutiny and we couldn’t decide what variables/bio-markers to test. There are a couple of efforts that have taken off since then to try to collect data on therapeutic modalities, including resveratrol. The largest-scale effort I know of is CureTogether but it isn’t very useful because the vast majority of reports appear to be subjective and unreliable (e.g. “I feel that resveratrol has slowed my aging …” and so forth). Such a web-based approach would be much more useful if objective tests like those you have done could be implemented but I’m skeptical you could get many people to produce and report data in an unbiased fashion. I have found that the desire to believe whatever you’re doing is good is incredibly strong and can be rationalized ad infinitum.
Interestingly, it looks like professional scientists and even big pharma might have gotten caught up in that mindset. Many of the reported benefits of resveratrol have been controversial from the beginning and recent reports suggest that neither scientists nor pharma can reproduce key results. Matt Kaeberlen, one of the first discoverers that sirtuin overexpression extends lifespan and co-founder of a biotech company with me in the early 2000s, returned to academia and has raised some red flags about the resveratrol research. He showed that the key assay used to discover resveratrol in a drug screen seems to depend on a biochemical artifact. Sirtris, a biotech company specializing in sirtuin research and that was bought by Glaxo for $720M, developed some resveratrol analogs that were reported to have multiple benefits, including control of type 2 diabetes. But recently Pfizer and Amgen have published studies saying they cannot reproduce Sirtris’s results. You can read many reports of this mess on the web but here are good, recent accounts of the controversy:
https://www.boston.com/business/healthcare/articles/2010/05/17/cambridge_biotechs_work_on_resveratrol_cuts_to_basics_of_biology/
https://www.nature.com/nbt/journal/v28/n3/full/nbt0310-185.html
https://www.newscientist.com/article/dn18396-stay-young-on-red-wine-drugs-think-again.html
https://seekingalpha.com/article/182123-what-did-glaxosmithkline-get-in-return-for-buying-sirtris
https://pipeline.corante.com/archives/2010/03/09/a_gsksirtris_wrapup.php
Well, as they say, the plural of anecdote isn’t data
yes, apparently the plural of data isn’t data.
yes, the New Yorker was a quite a downer last week.
I have a rare, fatal and incurable disease myself, and just this week a major trial finished with a negative result. No benefit found, despite promising preliminary results. Earlier this year another major trial failed, similar story. So far there have been dozens of treatments tried after earlier positive results, and only one has showed any benefit – and that only extends survival by two months.
I am a member of an online site called PatientsLikeMe which lets everyone record what treatments they are trying – supplements, experimental drugs and therapies, etc. You also record your rate of deterioration. Everyone can see everyone else’s data and the goal is to try to find common elements among people doing better than average. Unfortunately it doesn’t seem like anything works consistently.
It relates to the distinction between anecdotes and data. If one person said they felt a little better after trying, say, fish oil, that’s an anecdote. But you can go to PLM and see dozens of people who have tried it, when they started, what dose they used, and whether their progression changed. That starts to look like data.
I think the difference is that PLM is not a fish oil forum. That is just one of literally hundreds of different things people have tried. There is no reason to think there is a selection effect in reporting results, positive or negative.
Anecdotes are more likely to be biased. Someone who gets benefit from fish oil is more likely to talk about it than someone who noticed no change. This is why we are cautioned not to accept even multiple anecdotes as objective data. You need to get away from selection effects.
yeah, PatientsLikeMe is a great thing.
As I understand it, the non-reproducibility of results concerns the mechanism of action of resveratrol analogues, not the reported benefits of resveratrol itself.
Is there a global warming connection here?
The general pattern for new or fringe theories is that initially they are supported, and then many fall apart.
I think this is partly because initial studies are done by true believers, who may not be qualified researchers and who work in an undisciplined way on a small scale. Nevertheless, their results are published in the media and, nowadays, become conversation fodder on the internet.
Serious researchers avoid the topic at this stage. They don’t want to be stigmatized as cranks, and for controversial theories they don’t want the internet anonymous attack machine to come after them. But as evidence seems to be building up, they finally break down and start to look at things. This is when problems are found, and the initial theory starts to look flakey.
This process cycles over and over as the initial theory is modified. Because theories never, ever die, no matter how much evidence there is against them.
Marko, yeah, as stuff gets looked at more closely, the initial conclusions may not hold up. I guess that’s the connection with global warming.
Curious, too, is the assumption that all fish oil is the same.
How about trying bug oil? There are only a million or two species of bugs to try, but surely they’re all more or less alike. Next we can try bacteria. Bacteria must be good for us.
I think the resveratrol thing is mostly hype. I worked with Brian Kennedy and Matt Kaeberlein until just recently and they had been predicting for some time that the sirtuin assay was bunk– it was finally demonstrated in a recent study that Dr. Estep alluded to above. We used to crack jokes about resveratrol all the time in the lab because we thought the whole thing was ridiculous. It does seem to activate SIRT1 in vivo, but it’s not a direct effect as claimed. Resveratrol is certainly not a “calorie restriction mimetic”.
Any time high-impact publications, high-power labs, the media and private money are involved, you have to take results with a big grain of salt. Once the hysteria sets in, you get labs that pump out crappy studies that get published in Nature or Science without proper scrutiny. Then people can’t replicate the results, but they don’t publish that because who wants to contradict a Nature paper that everyone has already taken as fact? Either that or it gets published but it’s low-impact. So the grumblings usually remain in the dark corners of scientific meetings. This kind of intellectual pollution seems to travel along with hot topics like nutrition and aging. It was shocking to me when I first began reading diet-health papers.