Worms excel in life beyond menopause

Of the many species that outlive their fertility, lab worms (C elegans) take the cake.  They stop laying eggs after two days, but they can go on to live for two more weeks. Did they miss school the day that the teacher was lecturing on Darwinian fitness? Or perhaps do the elders of their community have a message for their grandchildren?

In the paper that launched the modern evolutionary theory of aging, George Williams [1957] predicted:

“(6) There should be little or no post-reproductive period in the normal life cycle of any species.” 

Not so long ago, evolutionary scientists thought that human menopause was unique in the biosphere. A whole industry arose explaining the “grandmother effect”: Playing the game they like best, population geneticists would try to concoct circumstances under which a woman’s “selfish genes” could leave more copies of themselves by not reproducing, but instead freeing their bearer to raise her grandchildren, while their mother went out hunting and gathering. 

Then a literature began to appear in which other animals were discovered to outlive their fertility. But this didn’t stop them. Seldom do facts on the ground discourage an enterprising evolutionary theorist.

In several whale species, the female undergoes menopause and continues to live for several decades thereafter [McAuliffe]. Elephants, lions, and baboons experience menopause [Packer], and all these animals might plausibly be argued to relate to their grandchildren.

But some other animals, not so much. As early as 2004, menopause was recognized as a “generalized mammalian trait”. Beyond mammals, hens and guppies both stop laying eggs. Yeast cells stop budding. 

James Vaupel’s research group published a comprehensive survey of animal “life histories” [2013], in which they charted fertility and mortality over a lifetime for dozens of species of plants and animals. The blue lines map fertility, and the grey background curves show survival.

In these 48 plots, C elegans worms, the common lab roundworm, stands out as living most of its life with no fertility at all. Why would that be?

My principal contribution to the evolutionary theory of aging has been to cite demographic stability. Population homeostasis doesn’t come for free; in fact all ecosystems would blow up if their constituent species were evolved to live and reproduce as copiously as possible. In general, aging has evolved as nature’s way to prevent population overshoot, to avoid extinctions that come when populations crash, and thus to keep ecosystems stable and productive in the long term. This is my theory paper, with Charles Goodnight, on post-reproductive lifespan as a demographic adaptation. 

The life cycle of C elegans in nature

C elegans worms are hermaphrodites. For the first day of their lives, they are male, and generate a few hundred sperm cells. Then they become female, and start laying eggs, one by one, fertilized by their own stored sperm. But after two days, they run out of sperm and can’t fertilize their own eggs. They become infertile, unless…

One worm in a thousand is not a hermaphrodite but a male. It lays no eggs, but produces oodles of sperm, which can fertilize the eggs of a hermaphrodite worm. But she has to be very lucky to find a male, because males are so rare. So this is an interesting footnote, but probably irrelevant to the reason worms are programmed to run out of sperm and discontinue fertility. The vast majority of hermaphrodite worms just live out their lackluster lives. Without sex, is life worth living?

When they are not growing in a Petri dish, C elegans worms live on rotting fruit in the wild. One piece of fruit on the ground can feed millions of worms over a period of several weeks. But how are they to get to the next piece of fruit?

If the worm is fat and happy, it lives 2 to 3 weeks. But if the worm is starved, it has a special adaptation. When food is scarce, it can go into a diapause stage called dauer, in which it doesn’t eat and its metabolic rate is very low. As a dauer, the worm can live for many months. The dauer can be ingested by a bird and survive digestion.

Carried by the wind or hitching a ride in a bird’s intestine, dauers are the worm’s plan for getting from one piece of fruit to the next. The “fitness” of a worm is probably better measured by its probability of finding a new piece of fruit than by the number of offspring it produces. Fitness is a property of a worm colony, not of a single worm. The fitness of a colony is related to the number of dauers that are sent out to explore the environment for the next piece of fruit.

Do the math — at several hundred worms per generation, the first three generations will do fine. They will be able to gorge themselves on fruit. But the fourth generation would be hundreds of times larger. The worms would start to grow, only to find that there are so many of them competing for the fruit that none have a chance to grow to maturity. 

This is why the ability to dauer is not just useful but essential to the worms’ collective survival. Larval worms detect food scarcity, and they flip a switch into dauer mode. They cast their future to the wind, or commit themselves to the gut of a hungry bird. 

There is a danger, though, that they might miss their timing. If they are starved right out of the egg, the worm is not yet large enough to form a dauer. But one day out of the egg, the worm is already committed to maturity, and it is too late to dauer. The dauering window is just a few hours long. It is easy to miss.

Here is where grandma can help. In a worm colony that is in its fourth generation, there are already hundreds of thousands of elder worms, post-reproductive but still squirming. These worms saturate the nest with dauer pheromone. The pheromone signals the hatchlings of the fourth generation that, even though it may seem that there’s plenty of food for now, the generation size is so high that the food supply is in danger of being exhausted before they can grow to maturity. The pheromone encourages the young larvae to go into dauer rather than take their chances on maturity and reproduction in present conditions. Dauer pheromone from the oldest worms is a signal both of crowding and of colony age. Exponential mathematics is what is threatening the worm population, and exponential mathematics is what potentiates the enormous increase in dauer pheromone that inform the larvae that their best prospects are to leave home and seek their fortune in the countryside.

This last paragraph is my personal theory, not yet fully confirmed. For a decade, I have been looking for someone to actually measure dauer pheromone from elder worm populations, and no one has taken up the challenge in this form. But from two labs where I have worked, there is partial confirmation. These are in the Beijing National Institute of Biological Sciences (Meng-qiu Dong and Wu Gang) and at Washington University of St Louis (Kerry Kornfeld and Andrea Scharf). 

Evidence for the grandmother worm theory

NIBS Beijing

In 2017, while I was a visiting scholar at the National Institute for Biological Sciences in Beijing, my host, Meng-qiu Dong, was kind enough to design two experiments with grad students Wu Gang and Peng Lei. 

First they raised infertile worms synchronized by age, so they could extract liquid from their growth medium when all the worms were a fixed age. They confirmed that dauer pheromone could be extracted from the medium, and that the quantity peaked for 5-day-old worms. Larvae grown in the medium where 5-day-old worms had lived were more likely to dauer than larvae in a control medium. 5-day-old worms are grandparents, but not yet great grandparents.

Second, they grew worms to old age on half a Petri dish, and then grew a fresh generation of worms on the same Petri dish. As the young worms moved around, they tended to avoid the half of the plate where the old worms had lived (compared to a control dish where no worms had previously grown).This suggests the worms are instinctively avoiding crowding.

Results written up informally here were never published because they are highly preliminary. I think the hypothesis warrants further experimentation.

Wash U St Louis

In the WashU lab of Kerry Kornfeld, my job was to model the worm’s physiology and life cycle based on individual measurements of their growth, food consumption, maturation, dauering, egg-laying, and death. My computer model aggregates the worms’ individual behavior to predict population growth and population crash in a test tube where food (bacteria) is added on any given schedule. In our first published paper, we reported good agreement between the model and population curves as measured with worm counting software. 

But there was one persistent disparity between the predicted population curve and the lab results. With daily feeding, the computer model predicts that population should reach a peak on the 10th day, before declining precipitously and then leveling off with smaller oscillations. But in the lab, the worm population peaked on the 13th day, a lower peak followed by a less calamitous decline.

Secretly, I was delighted. We had been working on this project since 2014, and I had predicted at the beginning that our most significant discovery would be finding that we could not accurately model the worm population without including pheromone communication in the model. I created and refined a computer model for the WashU team that had only individual worm behaviors. But on the side, for my own curiosity, I was adding a parameter that controlled worm communication mediated via pheromones. 

I tested the bare model (no pheromones) by varying parameters within the experimental range and beyond, but the time to the first population peak was very insensitive to all my parameters. This is because the population is increasing exponentially, so between the 10th and the 13th day, the population explodes by almost 100x. 

But one change in the model that is capable of reproducing lab results is to increase the probability of dauering as the population matures. This is indirect but tantalizing evidence that the old worms really are sending a chemical signal that warns the young larvae not to try to grow to maturity, but to dauer instead.

My theory and why it is significant

Chemical signals from the grandmother worms alert a late generation of larvae that they should not try to grow up, but go into suspended animation. This softens the incipient population crash and assures a bumper harvest of dauers that spread through the neighborhood in search of a new home.

My hypothesis is that populations are evolved for homeostasis, to avoid population overshoot and crashes that can lead to extinction. Roundworms in nature are subject to highly variable food availability, and they are highly adapted to protect against population crashes. 

The “selfish gene” model is still the prevailing paradigm in academic evolution. It is difficult to understand post-reproductive lifespan within this paradigm. If confirmed, this would be an example of an adaptation at the population level in which individuals cooperate to guard against extinction of the community. There is an established communication system, even in this primitive animal, explicitly for this collective purpose. 

Why not Spermidine?

Spermidine is found naturally in many foods. Blood levels decline with age. Supplementation extends lifespan and healthspan in various species, and lowers mortality in human trials. I have been unable to find toxicity data or any downside to spermidine supplementation, except that the effective dosage may be impractical.

Consistently across species, extra supply of spermidine prolongs the life span in an autophagy-dependent manner and counteracts age-associated pathologies such as cardiovascular disease, neurodegeneration, and cancer.”
— Guido Kroemer in Science [2018

If you haven’t followed Guido Kroemer, you’ll find he’s a wellspring of deep biochemical knowledge. 

Many life extension strategies involve hacking the body’s signaling systems, binding to specific proteins like SIRT or TOR. But spermidine is a small molecule, too simple to have specific signaling properties. Compared to even smaller molecules like H2O2 and NO, it persists in the metabolism long enough to be transported and affect multiple tissues. 

Autophagy is the recycling of damaged organelles (cell parts) and damaged biomolecules, a renewing process that has long been recognized as a necessary process for keeping the body young. Mitochondria are the organelles that generate energy, and ROS are their toxic byproduct. One of the most essential functions of autophagy is the recycling of mitochondria.

Fasting triggers autophagy, and this is part of the reason it is good for us. Longo’s Fasting-Mimicking Diet was designed to promote autophagy. The life extension and other benefits from spermidine are found to result principally from enhanced autophagy. (The way this is determined is to repeat life extension experiments with an autophagy gene disabled, and it is found that spermidine has no benefit [2016].) Spermidine was found to extend lifespan in yeast, worms, flies, and human cell cultures [2009].

Apoptosis=programmed cell death is an essential function for removing senescent and cancerous cells. We need apoptosis, and would be more vulnerable without it, but as we get older, apoptosis develops a “hair trigger”, and cells begin to commit suicide when they’re still healthy and useful. Overactive apoptosis is to blame for sarcopenia – the loss of muscle mass with age. Apoptosis is also implicated in the loss of brain cells that leads to Alzheimer’s Disease. Spermidine dials down apoptosis

Spermidine is an anti-inflammatory, and there is suggestion from animal models that it could relieve arthritis [2016]. This study found a stronger benefit for spermine, which is a metabolic product downstream of spermidine.

Foods with the highest concentrations of spermidine include wheat germ, natto and other soy products, mushrooms, some cheeses, and apples. An ounce of wheat germ or 8 oz of cheese contains 10 mg spermidine. Some of our spermidine comes from the gut microbiome. The probiotic Bifidobacterium LKM512 increased blood levels of spermidine in mice.

Spermidine dosage is curiously missing from some of the most important studies. 

In the mouse lifespan study that produced the best results [2017], mean lifespan was extended by 24%. Spermidine was added to the drinking water of test mice at a concentration of 3mM. Translating to humans is not straightforward, but here’s a rough estimate: If you drank 2 quarts a day of water or other fluids with this concentration of spermidine, you would be getting about 1 gram. To get this from food you would eat 6 pounds of wheat germ daily. The highest-dosage supplement pills I can find have 5 mg, so 1 gram per day = 200 pills. For some other supplements, pills of ½ g or a full g are not hard to find. I don’t know if spermidine is intrinsically expensive or if there is another reason that supplement companies gauge the dosage levels so low.

There are no human studies based on supplementation, but some authors have tried to correlate mortality with estimates of spermidine intake from (self-reported) food surveys. Again, they report results by dividing the test subjects into three groups, comparing highest and lowest (estimated) spermidine intake without specifying quantities [2018]. Mortality was reduced by about 25% in the tercile with highest spermidine intake, corresponding to about 3 years of extra life. Because the correlation is computed based on foods, this result may include confounding errors, as the foods highest in spermidine have other health benefits. 

This Japanese study from 2012 looks for factors that affect longevity across 48 Western countries. Because there are so many variables, the uncertainty is great, but dietary polyamines (including spermidine) stood out as a factor decreasing cardiovascular disease burden (r = -0.355, p = 0.007). 

In elders with mild cognitive impairment, there was an indication of possible memory benefits [2018, 2021]. Two studies suggest benefits for immunosenescence as well [2014, 2015]. 

Here’s an excellent short review, also by Kroemer [2019] as senior author. 

The bottom line

I don’t know why it has taken me so long to discover spermidine, or why it is not more frequently discussed by health gurus and biohackers. This month’s Life Extension magazine calls for trials of multiple combined interventions — an idea which I have advocated since 2018. Spermidine is on top of their list of substances to try in combinations.

I plan to self-experiment with high dosages, based on the very limited data from rodents.

I think I know why methylation clocks are failing

Methylation clocks promise to revolutionize testing of anti-aging interventions in humans, by measuring their effectiveness in a few months instead of having to wait years to see if they affect mortality statistics. But there are signs that the clocks we have can be deceptive. I think this is because the clocks capture aspects of aging that are defensive and aspects that are programmed, and there has been no effort to tease them apart. To the extent that the body is defending itself against perceived damage, the clocks are measuring the wrong thing. (This summary probably won’t make sense until you read the details below — please be patient.

I recently submitted an academic study on this subject.

We have a lot of life extension treatments that work in lab animals. Do any of these work in humans? In practice, we need aging clocks to test them, because testing them directly requires tens of thousands of people to be followed over decades for each intervention. I’ve been enthusiastic about aging clocks since Steve Horvath published his groundbreaking analysis in 2013. But recently I’ve realized that the field needs a course correction.

Epigenetic clocks are based on patterns of methylation in our DNA. Genes are turned on and off at different places in the body, at different times of day, and, crucially, at different stages of life. Methylation is the most accessible and easiest to measure of many methods by which the body turns genes on and off. By focusing on the methylation patterns that change consistently over a lifetime, Hannum, Horvath and others who came after them have created computer algorithms that can calculate a “biological age”.  

The idea of “biological age” doesn’t depend on your theory of aging. But the utility of epigenetic clocks in assessing the benefits of putative anti-aging measures certainly does depend on fundamental concepts about aging, about which experts are still divided.

Why does gene expression change in old age?

“Nothing in biology makes sense except in the light of evolution.”
— T. Dobzhansky

1)    If you believe in programmed aging, then the directed changes in gene expression are means of self-destruction. Genes are turned on that increase inflammation, destroying arteries and neurons. Apoptosis is up-regulated to the point where healthy muscle and brain cells are dying. Protective anti-oxidants, DNA repair, and autophagy are down-regulated. All this destruction is accomplished via turning genes on and off. If any intervention sets back the methylation clock, then there is less self-destruction, more repair and maintenance. We expect that the body will live longer if the methylation clock reads a lower age.

2)    If you believe the neo-Darwinist theory that the body cannot be purposely destroying itself, then aging is an accumulation of incidental damage at the cellular and molecular levels. If there are associated epigenetic changes, these cannot be causing the destruction, so they must be a response to the damage. Changes in gene expression as captured in the methylation clocks must be the body’s effort to protect itself with increased immune function, increased autophagy, increased antioxidants, increased DNA repair. If any intervention sets back the methylation clock, then there is less repair and maintenance. We expect that setting the aging clock back to a younger age will actually decrease life expectancy. This insight is counter-intuitive, but, if correct, it changes the logic of methylation clocks. 

For people who don’t believe that aging is an evolved program, the whole idea of a methylation clock is a non-starter. No matter how accurate the clock is, setting it back is counter-productive. Even if the clock is calibrated to markers of health (like the PhenoAge clock) or calibrated to actual mortality (GrimAge), it is still based on the body’s response to damage, and not on the damage itself. Setting back the clock is counter-productive, because it means dialing down the body’s repair and maintenance system. 

Since 2013, there has been a kind of double-think in the world of anti-aging research. Most researchers, at least in public, continue to embrace perspective (2), even as they adopt methylation clocks to evaluate the interventions they develop.

All this is assuming perspective (2). But I’m notorious for being a proponent of perspective (1). From perspective (1), turning back the methylation clock is a good thing. It means that the body’s program of self-destruction is dialed back. So where’s the problem?

In recent years, I have become convinced that epigenetic changes of both types (1) and (2) are taking place simultaneously as the body ages. The body is at war with itself. The self-destructive adaptations listed above are real: dialing down repair and maintenance, promoting systemic inflammation, apoptosis of healthy cells, derangement of the immune system. But the body retains its protective responses, and there are also changes in gene expression that ramp up the repair processes. All the present clocks include a mixture of (1) and (2); this is why we do not yet have a reliable metric for the efficacy of anti-aging technologies. 

What is the evidence that changes of types (1) and (2) are both components of all extant aging clocks?

Some of the best-established interventions for extending lifespan do not affect the major algorithmic clocks, or do so modestly compared to what might be expected from their observed effects on lifespan. Rapamycin extends lifespan of male mice without affecting their methylation age in the Horvath rodent clock. Participants in the CALERIE study who have adopted extreme CR diets showed no significant benefit according to either the GrimAge or PhenoAge clocks.

Conversely, Katcher’s intravenous infusion of exosomes (E5) has a dramatic effect on the Horvath rodent/human clock, reducing epigenetic age by half, but thus far seems to extend lifespan less than the clock setback would imply. The Conboys recently published a withering criticism of the utility of current methylation clocks, and of the machine learning algorithms from which they are created. They report that clocks in common use do not respond as expected to known life-shortening conditions, such as Down Syndrome, inflammaging associated with arthritis, and Parkinson Disease.

Here’s what first clued me in

The GrimAge clock of Lu and Horvath was trained on actual mortality data, using historic blood samples for which the future history of the donors was known. This was a major advance from previous clocks. But one element of the GrimAge development alerted me to the issue concerning type (2) changes, as described above.

Part of the training of GrimAge involved a methylation image of the subject’s smoking history. Smoking is known to accelerate aging and shorten life expectancy. Certain patterns of methylation are associated with smoking, and are also valuable predictors of time until death. These were included in the GrimAge algorithm.

My assumption was that smoking decreases longevity by damaging tissue of the lungs, not by turning on the program of self-destruction. Therefore, if there are methylation changes associated with smoking, they are probably of type (2). In other words, the methylation signature of a smoker who scores as “older” in GrimAge is likely to include activation of more protective pathways than a non-smoker who scores “younger”.

This is an important clue. The methylation profile of a smoker is useful in constructing a GrimAge clock, but it should be counted in reverse. Methylation changes associated with smoking are statistically associated with shorter lifespan, but mechanistically with protection. These changes should have been included in algorithmic clocks with negative coefficients, signaling a younger biological age. This was not how the GrimAge clock was constructed in fact. Methylation changes associated with smoking were included in the GrimAge clock with positive coefficients (simply because they are statistically associated with a shorter life expectancy).

In general, the methylation image of smokers is an example of type (2). All type (2) changes should be counted with negative coefficients in methylation clocks, even though they are statistically associated with older ages and shorter remaining life expectancy.

So it is crucial to distinguish epigenetic changes of type (1) from type (2)

 The story of GrimAge carries a message that suggests ways that methylation changes of types (1) and (2) might be teased apart in algorithmic clocks. Present clocks don’t distinguish between (1) and (2) so presumably the two types of methylation changes are combined in a way we might connote as (1) + (2). The goal would be to create a clock built on type (1) changes alone, or, more speculatively, penalize the clock for type (2) changes, so with the result that the algorithm measures (1) – (2) rather than (1) + (2).

The long-term goal would be to understand the metabolic consequences of each CpG change, separately and in combination, so that a clock could be constructed with full confidence that it scores beneficial and detrimental methylation changes appropriately.

Lacking this understanding in the interim, we might make progress toward distinguishing (1) and (2), by learning from the smoking example. One way to acquire a database of type (2) changes is that animal models might be injected with pro-inflammatory cytokines, and their epigenetic consequences mapped. Since the inflammation is imposed externally, we should presume that the response is all type (2). 

Similarly, the animals’ immune systems might be challenged, or they might be subjected to laceration or small doses of radiation, again to chart the epigenetic response to compile a list of candidates for type (2) changes. These experiments could not ethically be performed on humans, however there are humans whose aging is accelerated by non-epigenetic factors, including alcohol and drug abuse. Such people might be tested as part of the quest for type (2) changes. People healing from physical and emotional trauma might also be presumed to have epigenomes modified in the direction of type (2).

Hormesis is the body’s overcompensation to challenges. The body is damaged by something we do or we eat or suffer from, but the body overcompensates to the damage such that we live longer. 

Caloric restriction (CR) is the best-established example of hormesis. We might have most confidence in the epigenomes of people and animals subjected to caloric restriction]. Across the animal kingdom, CR is the most robust anti-aging strategy known at present, and we can be confident in subtracting CR-associated epigenetic changes from any algorithmic measure of biological age. These changes can be observed in humans, as in the CALERIE study mentioned above; or they can be observed in rodents, which have enough commonality to human metabolism that some of the same methylation sites have common functions in both. Our methylation clocks should be calibrated to be sure that the changes associated with CR are scored toward a younger biological age. 

In addition to CR, there are dozens of interventions known experimentally to extend lifespan in rodents, including juvenile exosomes, rapamycin, certain peptides, vitamin D, NAC, certain anti-inflammatories and angiotensin inhibitors. Recently, some of these have been tested for their effect on algorithmic clocks; and this has been interpreted as evidence for or against the intervention. We might reverse the logic and interpret the same data as training to calibrate clocks, assuming that these changes must be beneficial, and the clock algorithm should reward them with a younger age. If an intervention is known to increase lifespan, then we may presume that epigenetic changes observed in response to that intervention are beneficial.

Before 2013, biological age was estimated with measures of performance and appearance: grip strength, gait speed, athletic endurance, memory, exhalation volume (FEV), skin wrinkles, arterial inflammation, cartilage integrity. In the age of epigenetics, these physical characteristics retain their value as predictors of mortality, and a hybrid clock might be devised, combining physical and epigenetic factors. 

For the future

I know of projects at Stanford (the Biomarkers of Aging Consortium), at Tally Health, and at Tru Diagnostic to develop the next generation of clocks to evaluate anti-aging interventions. There are probably other, parallel efforts that I don’t know about, but I will hear about them in your comments. I hope that the theory and suggestions in this blog may be useful to them.

Michael Levin on Aging and Regeneration

The Levin laboratory has found the developmental blueprint that tells the embryo how it must grow. The language of morphology is not genetic or epigenetic, not biochemical at all, but bioelectric. If, as I and others have speculated, aging is an extension of development into a program of self-destruction, then maybe aging follows a bioelectric blueprint. Maybe electrical interventions have a role in anti-aging therapies, supplementing or replacing the biochemical interventions that have dominated research in gerontology. Below I review a prospective analysis from the Levin lab.

Background — Levin’s laboratory has put bioelectricity on the map

Michael Levin is my favorite living scientist. He is firmly rooted in laboratory biology, yet he has devised and interpreted his experiments so as to cast new theoretical light on biological mechanisms, again and again.

Levin studies morphology first. Why do bodies take the form that they do? How does the developing body know which organ to grow in which location? How does a hand know to generate exactly five fingers? How does the limb know when it is the right size, and it’s time to stop growing?

His signature thesis is that embryonic development is guided by electrical patterns, rather than gene expression. There are patterns of electrical potential that are maintained by cell walls that selectively allow the passage of some ions but not others. Voltage patterns imprinted across embryos contain detailed information about the shape and form of the adult, and how the embryo is to develop toward that goal.

At this point, you are wondering where the electrical patterns come from. The egg and sperm didn’t have these imprinted patterns, and the only information they carried was genetic and epigenetic. This is a crucial question, as Levin acknowledges, but he doesn’t yet have an answer.

Levin’s favorite model animal is the flatworm, Planaria. These worms rarely go through the trouble of sexual reproduction (which is essential in the long term for preserving the integrity of the genome). Commonly they reproduce clonally. Any small piece of a worm can regenerate a full body. As a result of long periods of asexual reproduction, the genomes of these worms are a mess, with different chromosomes — even different numbers of chromosomes — in a mosaic through the body. But no matter! The worms seem to develop their adult shape and function fully independent of what selection of genes happen to be available.

In one of Levin’s signature experiments, he is able to modify the electrical pattern in a piece of a worm, not touching any of the biochemistry, and the result is a worm with two heads and no tail. Or two tails and no head. Both seem to be perfectly viable, and survive well, at least in the lab environment. The two-headed worm breeds true, in that pieces of a two-headed worm will regenerate another two-headed worm, and pieces of a two-tailed worm will regenerate another two-tailed worm. But if the experimenter artificially induces a change in the electrical pattern, he can cause the progeny of either of these to revert to the boring phenotype with one tail and one head. The two-headed worms and the no-head worms do not differ genetically from the normal type.

I’m in danger of getting carried away with these results before I even get to Levin’s paper on aging. I’ll just mention two more.

Learning. You can train a worm in simple behaviors. Then cut off its head, and the tail grows a new head that remembers the behavior you taught to the old head. This is one of several counter-examples to the universally accepted common sense that memory lives in the brain. (The experiment was repeated in Levin’s lab, but originated decades ago.)

Adaptation. Barium salts poison the ion channels that cells use to maintain their electric potentials, and in planaria they cause a worm’s head to shrivel up. The worm grows a new head, and it is resistant to barium salts. Levin notes how unexpected this is. Worms in nature never encounter high enough concentrations of barium to cause a problem, so this is not an evolved adaptation to barium. And yet it is immediate. The worms know to turn on a handful of genes that are otherwise silenced and it solves their problem.

the cells detect (and act on) highly processed state information several steps removed from the proximal events at the membrane. In this case, the many ways to depolarize tissue could be naturally coarse-grained to represent a single problem: a change in membrane voltage addressable by a single set of transcriptional actions… Acting on such coarse-grained information is a simple form of meta-processing (i.e., “higher-level” information processing that controls some lower-level process). [ref]


Most of the paper is reviewing relationships between electricity and aging. The list is impressive, but perhaps this is because electricity is the unsung hero of many biological processes. The last 50 years of bioscience have been focused so intensively on the chemistry of DNA and proteins that we have a distorted idea of their importance. Life depends on electric forces within and between cells. Nerves operate via pumping Ca and Mg ions across membranes to create electrical signals. Exchange between Na and K ions causes muscle cells to contract. Mitochondria provide energy to all our cells by converting chemical energy to electric energy, The zeta potential is a charge on blood cells moving through our arteries, while preventing clotting.

Some of the relationships of bioelectricity to aging, all of which were new to me:

  • Rejuvenation via exosomes was shown to work (among other mechanisms) via modification of K+ and Ca++ ion channels
  • Oxytocin, identified by the Conboys as a key anti-aging blood factor, works by regulating depolarization and firing of neurons.
  • Rapamycin and resveratrol both inhibit voltage-gated K+ channels
  • Quercetin and other senolytics regulate ion channels
  • Unsurprisingly, brain aging is intimately linked to Ca++  ion pumps, which form the basis of neuronal signaling. Calcium channel blockers can protect against Parkinson’s.
  • The heart is dependent on bioelectric function both to run its clock and to fire the pump muscles. Levin lists some ways that CV aging is connected to bioelectricity.
  • Cancer is a special case…


Levin’s lab has rediscovered a new property of cancer cells that came to light only late in the game, and still occupies a backwater of cancer research.

“It has been known for decades that tumorigenesis begins with the bioelectric decoupling of cells from the somatic morphogenetic network (2003; 2010; 2006).”

Normal cells are in electrical communication with all their neighbors. But cancer cells cut off all communication, and each cell is isolated from information about its surroundings. Levin was able to revert cancer cells to well-behaved growth simply by restoring their electrical contact with neighbors.

The old theory about the etiology of cancer is that it starts in the nucleus, with a series of compounding mutations. Schaefer and Israel disproved this, to my mind, by transplanting mutated nuclei of cancer cells into normal cells (they remained normal cells) and transplanting healthy nuclei into cancer cells (they remained cancer cells). Despite this clear demonstration almost 30 years ago, most oncologists still believe that cancer is caused by mutations, rather than that mutations are caused by something outside the cell nucleus. Schaefer and Israel seemed to show that the the real source of cancer is in the cytoplasm. But their work is also consistent with Levin’s hypothesis — cancer is all about electrical communication networks.

How are adjacent cells connected?

Cells are enclosed by their individual cell membranes, and within these membranes are embedded proteins called connexins that can hook into connexins from an adjacent cell, forming a bridge. The bridge conducts charged ions and small molecules across the gap junction, leaving large molecules on either side.

(This sounds like synapses connecting nerve cells. Synapses have gap junctions, but their gap junction channels do not involve connexins. Levin reminds us frequently that the electrical networks formed by neurons are just one manifestation of a ubiquitous and much older phenomenon of electrical communication between animal cells.)

How is this useful for cancer therapies? Levin has experience using drugs to open and close ion channels between cells and create a desired pattern of electric potential in a group of cells. He has done it successfully for years in planaria. Can the same technique be used to recreate healthy relationships among tumor cells, redirecting the transformed cells to become functioning team players in the body? The part that has already been accomplished (using voltage reporter dyes) is to measure electrical patterns. The challenge, as Levin reports, is to create computer models which will predict the effect on these patterns from various combinations of ionophores.

Another, simpler idea is immediately available. Animal cells almost always carry a negative charge, maintained by assorted ion pumps in the cell membrane. Cancer cells tend to collect positive Na+ ions and lose their negative charge. The simple remedy is to deliver an ionophore drug that tends to pump positive ions out of the cell. Preliminary experiments suggest this has both a preventive effect and could be a treatment for cancers that have already developed [ref]. Bioelectric drugs were shown to have anti-cancer effects in cell cultures of human glioblastoma [ref].


“Regeneration and longevity are intrinsically linked”

This idea is the underlying motivation for Levin’s foray into the science of aging. Evolutionary theory would suggest that animals that have long gestation times, slow development, and gradual reproduction have greater investments in each individual and a greater need for a capacity to regenerate; but this does not imply that loss of regenerative potential is related to the root of aging. Salamanders are the best example of the correlation; they have extraordinary ability to regenerate, and they have extraordinary lifetimes for their size, up to 30 years. Planaria don’t age at all, and their capacity to regenerate is even more impressive than salamanders. Sea stars also have remarkable regenerative capacity, and their lifetimes are long, but less impressive than salamanders. On the other hand, octopuses can regenerate entire legs, but their lifespans are short for their size. Giant octopuses over 100 pounds only live 5 years.


Regeneration sounds attractive — but does it address aging?

Here’s a factoid I remember reading more than two decades ago, but in all that time, I’ve never been able to rediscover the reference. Sea stars have regeneration ability comparable to planaria — you can cut them up in small pieces, and each piece generates a whole new sea star. But unlike planaria, sea stars reproduce sexually and they age gradually, with a fixed lifespan of about 8 years. If you cut off a sea star leg, it will regenerate the whole body, but the body will remember the age of the original animal and die on the same schedule. A weaker example of which I’m more certain is the octopus. Octopuses have limited regeneration ability, but they can grow back one or several severed tentacles. Octopuses are semelparous, and the fact that they grow back new tissue doesn’t change the fact that they will die following reproduction.

The relationship between regeneration and aging is largely unexplored. The Levin lab is way out in front studying regeneration. It will be interesting to see if aging can be addressed with the ionophores and other “morphogenetic” tools that they use to modify voltage patterns in animal tissues.

Programmed aging — where is the program?

Approaches to anti-aging medicine have been divided between those that work at the cellular level (bottom up) and those that work at the systemic level (top down). You probably know that I have been an advocate for top-down from the beginning. The body knows how to repair itself, and just needs the orders to do so.

I have imagined those orders to come from signal molecules in the blood — hormones and cytokines and ribozymes = active RNA. Levin’s lab has elucidated another level of centralized control using electrical signaling. Aging is programmed. To what extent is aging programmed chemically and to what extent electrically? It’s a question to be explored with new experimental paradigms.

You’ll forgive me, knowing who I am, if I point out that the weakest part of this paper is that it considers only the loss of electrical information, and never the possibility that electrical information could be purposefully modified in a manner that is destructive to the integrity of the organism. For example, the use of the word “corrupted” shows that the authors are thinking in terms of lost information rather than deliberate rewriting of the pattern.

A key question for the biomedical applications of morphoceuticals for aging is how the bioelectrical pattern is corrupted over time and if an enhancement/repair of it will lead to proper maintenance or rejuvenation.

The introduction pays lip service to “programmed aging” as one of the viable theoretical frameworks, but this idea is clearly not digested. References cited for “programmed aging” are all for authors who are hostile to the idea. My book is not cited, nor are cogent accounts of the experimental case for programmed aging by people with reputations far greater than mine [ref, ref, ref, ref].

I find it interesting that even though Levin is always thinking in terms of logic circuits and electric controls, the tools he uses are always chemical. Drugs are used to move ions around and modify electric patterns. No one, not even the Levin lab, has found ways to modify the patterns of electrical potential directly.

Programmed aging requires a reference time, localized or spread through the body, a repository of information about age. If we could find the clock and set it back, like rewinding the odometer of a car, the body would make the appropriate repairs and adjustments to recreate a young body.

I have speculated the clock is in the hypothalamus, where chemistry and electrical memories meet; or maybe it is distributed in the epigenetic state of cells throughout the body. But how does it keep time? Is it logically possible for the clock to be homeostatic without some external reference time? Are there multiple clocks that consult each other and find a consensus (what Aubrey calls ‘“crosstalk”)?

Is one of these age clocks, at the highest level of control, based on patterns of bioelectricity?

Dr Mercola Doesn’t Like Seed Oils

I’ve learned a great deal from Dr Mercola over the years. His orientation toward natural approaches to long-term health, using diet and lifestyle in preference to pharmaceuticals, aligns with my own. I agree with most of what he has to say.

So, when he launched a campaign vilifying seed oils in the diet, I stood up and paid attention. “Linoleic acid found in vegetables and seed oils may be the biggest contributor to chronic disease in the Western world.” [from Dr Mercola’s Censored Library]

Last spring, he published a review article on health hazards from Ω 6 fatty acids in general and linoleic acid (LA) in particular. He makes a multi-pronged theoretical case linking LA to defective cell membranes, leading to diabetes, cognitive impairment, heart disease, and cancer. At the end, he tries to tie these concerns to real life results with epidemiological studies. To me, it seems that the theoretical arguments are strong, but epidemiology doesn’t seem to support his fears. In particular, most nuts are abundant sources of LA, and yet nut consumption is robustly linked to good health and long life.

Chemistry background

Petroleum oil consists of chains of carbon atoms surrounded by hydrogen. These have no biological presence. Saturated fats (biological oils) in biology are similar chains with COOH at one end — essentially petroleum-like molecules with vinegar at one end. Unsaturated fats have a double bond between carbons, which is a place where there are missing hydrogens, and an angle or kink in the chain.



Saturated fat: 

Unsaturated fat:


For omega 3s, the kink is close to the end opposite the vinegar end. For omega 6s, the kink is closer to the vinegar end.


(The pictures are alpha linoleic acid, ALA, an omega 3 (left), and and gamma linoleic acid, GLA, an omega 6 (right).)


Petroleum oils are biologically inert because there are no enzymes that can attack the carbon-carbon and carbon-hydrogen bonds. But the vinegar end of a fatty acid is a biochemically active site, and each double bond in the fatty acid is another active site. 

Food Chemistry

Animal fats tend to be saturated (no double bonds). They are more viscous. Vegetable oils end to be unsaturated, less viscous, more chemically active.

The double bonds make it easier for the body to work with a fat and put it to use. But each double bond is also a place where oxygen can attack the molecule and turn it into a useless and harmful product. OXLAM means “oxidated linoleic acid metabolite”, and it is not a single chemical, but a class of chemical byproducts which, Mercola says, can cause disease.

Even before it enters the body, unsaturated fats are prone to being oxidized in high-temperature frying. Saturated fats are theoretically better for frying because the double bonds in unsaturated fats are target sites for oxidation. So, for home-fried foods, it is safer to use coconut oil, ghee, butter, or lard. Restaurant and commercially fried foods are often made with unsaturated seed oils, which are likely to be oxidized by the time we ingest them.

The benefits of maintaining the proper omega 3:6 ratio are well-established. Bodily tissues consist primarily of saturated and monounsaturated fats, which are a readily utilized source of nutrients that support the development and maintenance of cells. The primary PUFAs are omega-3 and omega-6 fats, which the body needs in relatively small quantities [ref].

It is important to consume sufficient amounts of omega 3 fatty acids to sustain optimal health, with the recommended daily serving being between 500 and 1000 milligrams of omega 3 [ref, ref]. Contrary to previous perceptions, however, consuming larger servings of omega 3 fatty acids does not support an ideal ratio. Instead, excessive quantities of omega 3s may cause additional metabolic damage—similar to that which occurs due to the conversion of elevated LA levels. [from Mercola & D’Adamo]

It is well accepted that omega 3 fatty acids have benefits for reducing inflammation. Mercola claims that animal-derived omega 3’s are better in this regard, on the authority of this review. DHA and EPA are long-chain omega 3s, and are converted by our bodies into resolvins that cool inflammation. They come from either fish or ocean algae sources. ALA is a short-chain omega 3 that comes from walnuts, chia seeds, flax seeds and other vegetarian sources. ALA can be converted to DHA in the body, but slowly and inefficiently.

Examine.com  agrees

Omega 3 fatty acids in flax seed (as well as in Hemp Protein) are found in the form of Alpha-Linolenic Acid (ALA). Not only is ALA not sufficient to supplement on its own[1], but ALA has to be converted by the body into a usable form, and the ratio of conversion from unusable form to usable is rather poor, somewhere in the range of 5-15%[2]. Omega 3 supplements in the form of EPA and DHA are what the body tends to use for many of the benefits associated with fish oil.

For vegetarians and vegans, supplementing with DHA from algae can “markedly enhance the DHA status (of serum and platelets)” and “provide for the formation of substantial EPA”[3]. Supplementation of ALA and/or GLA is not enough[4].


Mercola fingers inadvertent oxidation as the process that turns LA into toxic OXLAMs. LA becomes incorporated into cell membranes, where it can remain for years. The membranes are then more fragile, and oxidation of LA in the membranes makes them porous, not functioning as the cell needs them to function. A particular OXLAM called 4-hydroxynonenal (4-HNE) accumulates, and high levels can trigger programmed cell death in the same way that peroxide is designed to do. 

Cardiolipin, created from four fatty acids, is essential for the efficient operation of mitochondria. Mitochondria are particularly vulnerable to oxidative damage, and incorporation of LA into cardiolipin is problematic for the mitochondria. Misformed cardiolipin distorts the convoluted shape of mitochondria, impairing their function

Health statistics

The epidemiological evidence for Mercola’s worries is weak, and he admits as much. There are conflicting studies, associating LA level positively and negatively with CV risk. Mercola claims the same is true of LA and diabetes, but the beneficial effects are documented by epidemiology, while the paper he cites for detrimental effects is theoretical. The effect of LA on cancer is mixed, and too small to measure

In this study, higher levels of circulating LA were associated with slightly lower all-cause mortality. In this study, addition of safflower oil to the diet was associated with slightly higher all-cause mortality.

Nuts — a counter-example

Nuts generally have high fat content, with 80% or more of the total calories coming from fat. Most of that fat is unsaturated (87% in peanuts, 95% in almonds, 80% in cashews), so Dr Mercola recommends limiting nuts to a small part of the diet. But many epidemiological studies have associated nuts in the diet with modest but significant benefits for CV risk, insulin resistance, and all-cause mortality,. [meta-analysis, review]


This recent review highlighted the benefits of omega 3s for brain health. The focus is on long-chain omega 3s, derived from either fish or algae. Plasmalogens are singled out as a promising supplement for preventing and even reversing dementia. Plasmalogens are ether phospholipids constructed by the body out of ingested fatty acids including ALA. Plasmalogens are essential components of cell membranes, especially in the nervous system. Levels decline after age 40, and declining plasmalogen levels are associated with cognitive decline. 

Confusingly, the word plasmalogen in the singular is sometimes used to denote a particular short-chain polyunsaturated species, AKA plasmenyl-phosphatidylethanolamine. Dietary supplements of plasmalogen are derived from cow milk or soy, and have been applied successfully in dementia trials. 

Dietary implications

Though there are many things that Mercola gets right, he has never understood aging as internally signaled self-destruction, so he is more fearful of accumulated damage than I think is appropriate. Oxidized lipids fits well with his damage model of aging. I understand the theory, but I think the real-world evidence of epidemiology is more compelling, and epidemiology tells us that nuts and even olive oil are pro-longevity. 

I agree with Mercola that the story about saturated fats ⇒ cholesterol ⇒ cardiovascular risk is discredited. Nevertheless, I’m a vegetarian for 50 years now. There is some evidence that vegetarians live longer, but my personal motivation is more empathetic than scientific [newer ref].  

I continue to eat a lot of nuts in the context of a low-carb vegetarian diet. I make salad dressing from olive oil and avocados, and cook with coconut oil. In the past, I have gone out of my way to include chia seeds and flax seeds, both for fiber and for Ω3s. Having learned what I did in writing this piece, I will back off the chia and flax seeds because their short chain Ω3s do not offer significant value. I don’t eat fish, so I will continue to supplement generously with fish oil and krill oil. It’s possible to get Ω3s from vegetarian ocean sources, but it’s expensive.

I’ve followed this diet for decades and it works for me, but your metabolism is different from mine, and diets are individual by nature. There is no one optimal diet.

I’m not going to write about the British mortality data

It would be just like me to tease you with a clever, ironic headline like this introducing a long, detailed analysis of the British mortality data. But in this case, I’m being serious. This article is about why I think the data are unreliable and unworthy of our credence.

In August, the British Office of National Statistics (ONS) released data on all-cause mortality broken down by age and vaccination status. The ONS data covers months from April, 2021 through May, 2023.

This is exactly what we need to determine if the COVID vaccines have saved lives. Data by “cause of death” are tainted because hospitals have been incentivized to treat deaths “with COVID” differently from other deaths. But total death counts are harder to corrupt. I wish that more countries would release such data. (Are you listening, CDC?)

I’ve written about dangers of the COVID shots, including one analysis estimating deaths in the hundreds of thousands for the US alone. I should be eager to vindicate my prediction with new data, and the aggregated British data appear damning.

On its face, the straightforward, aggregate calculation says that vaccinated people are three times more likely to die than unvaccinated.

The British data is broken down into seven age categories. (Children are not included. The lowest age category is 18-39, and thereafter the age categories are by decade.) Let’s zoom in on the youngest age group:

A small number of subjects had bad reactions to the first shot, and discontinued from there. These people have elevated risk from the vaccine, but among those who went on to get the full sequence + boosters, the shots appear to be protective.

Other age groups show the same pattern. For example,

For every age group, multiple COVID shots appear to be protective against death from any cause.

A statistics lesson

Of course, it’s mathematically impossible that within every age group, the vaccinated are at lower risk of death but when you aggregate all the age groups together, it turns out that vaccinated people are at a higher risk of death.

That’s what I thought. I spent a whole evening checking and rechecking the calculation.

But I might have remembered Simpson’s Paradox.

The data are correct. The graphs are correct. The key to understanding the situation is that the lion’s share of the deaths (the numerator) are in the oldest age groups, but most of the head count (the denominator) is in the younger age groups.

Crucially, the vaccination rate among the young is much lower than among the old.

So what we are seeing in the aggregate chart is that the oldest people are more likely to be vaccinated, and they are also more likely to die. But that doesn’t mean that vaccination is causing the deaths.

So, does this vindicate the vaccines?

I have written and others have written that side effects from the COVID vaccines, including risk of death, are so much higher than vaccines of the past that it’s hard to see them on the same scale.

Furthermore, there is insurance data showing that death rates are up as much as 40% in working people, and anecdotes tell of people having sudden heart attacks and (widely “debunked”) turbo cancers.

How can this be reconciled with the idea that, in each age group, multiple COVID vaccinations are associated with lower rates of death from all causes?

In the 18-39 age group, almost no one is dying from COVID. The most significant causes of death are accidents, suicide, and overdoses. It is not reasonable to think that three COVID shots have the effect of lowering these risks by 20%.

The most plausible explanation is a strong “healthy user” effect. People who are unvaccinated are, on average, poorer and less educated than the rest of the population, and it is well-accepted that death rates are closely correlated with social class, especially at the lower end. More speculative is that the data has been tampered with, deliberately doctored to hide the vaccines’ dangers. See below, “Fraudulent manipulation…”

For these reasons, I think there is not much to be learned from the new British data.

A pattern too stark to ignore

There is one result embedded in the data that is highly significant (p<0.00002) and implies a role in the vaccines for excess mortality in 2021. During the months March – December, the all-cause death rate in the UK rose by 30%, and the rise was tightly correlated with the cumulative number of vaccinations (R=0.96).

This rise did not continue into 2022, but the overall death rate remained elevated. This pattern, again, cannot be explained as simply as a huge death toll from the vaccine. Thirty percent is enormous, and cannot be hidden by any government’s manipulation of the statistics.

I know… Ed Dowd has famously documented a 40% rise in insurance claims against corporate group life policies; but it was only possible to hide that 40% bump because the base rate among these young people employed by large corporations was so low that 40% of that was too small to show up in a graph of whole-population mortality, which is always dominated by the oldest segments.

Since the British mortality increase is in the population as a whole, it must be concentrated in the older population. Part of the 30% rise was seasonal. Death rates increase 25% winter over summer. But part of the increase is likely due to a lingering effect of the vaccines on mortality and morbidity in the oldest group. Evidence for this is in the continuation of the graph into 2022 and 2023. Death rates are more variable, but we see that they reach new highs in the winter of 2023 and never return to baseline.

Fraudulent manipulation of the ONS data?

Last winter, Madhava Setty publicized research by Norman Fenton suggesting that ONS had tampered with their COVID data. The strongest evidence was for the early months, Jan-Mar of 2021, which are unaccountably absent from the latest edition of the ONS data. The current edition begins with April, 2021. The following graphs are derived from the ONS data release in July, 2022.

The charge that is easiest to verify is that deaths from the vaccine rollout have been shifted into the “unvaccinated” column. We can see this in the form of a spike in deaths of the unvaccinated just as the vaccines were being rolled out. This is most prominent in the death rate for the oldest groups.

This spike is hard to miss — three times as many deaths as expected in the small “unvaccinated” group in those first months when the much larger vaccinated group of elders was being given early priority for the vaccine. Nursing homes were first to be vaccinated.

In the spring, the vaccine was distributed to middle-aged citizens, and — remarkably — we see a similar bump in deaths of “unvaccinated” people in that age group. The bump started later and is more spread out, corresponding to the fact that middle-aged people signed up for the vaccine more gradually.

Starting in the summer, the vaccine was recommended for younger people, and we see a bump for mortality in “unvaccinated” adults under 40 begin in June.

The Jan-Mar data for 2021 has been censored in the new ONS release, and with it the evidence for the obvious spike in the older age groups. But what of the humps for younger age groups? Are they visible in the new data?

To answer this, I’ve plotted data from 2022 and 2023 on the same graph for each age range, so we can see where they match and where they don’t. At older ages, the data match nicely.

But for younger ages, the data have been revised upward, making the humps more prominent. I interpret this to mean that more post-vaccine deaths have been moved into the “unvaccinated” group.

For the youngest ages, it looks as though the data were adjusted more, proportionally, but the scale is much smaller because the death counts are very low for the young.

The bottom line

I’ll end where I began this entry. I’m not going to analyze the British ONS data because I don’t think the numbers are reliable.

News from Harold Katcher’s Lab

The big news: (1) the active ingredient in E5 is exosomes. (2) Young exosomes retain their ability to rejuvenate across mammalian species lines. Work that remains: (1) calibrate a new generation of methylation clocks. (2) Combine removal of old exosomes with addition of young exosomes. Optimize dosage and timing of treatments.

What is in E5, Katcher’s fountain of youth? From the beginning, it was described as a “plasma fraction”, leaving us guessing. Blood plasma contains thousands of chemical species in tiny quantities that nevertheless have powerful, systemic signaling effects. I thought once the patent was applied for, there would be no more need for secrecy, and the patent would reveal the formula; but the patent application was for a process, not a substance. Katcher and Sanghavi claimed rights for a broad array of substances extracted from the blood in different “embodiments” of the patented technique.

Then, in last month’s preprint posted on BioRxiv, the authors casually revealed the secret, almost parenthetically. “the exosome fraction of the plasma, which we term as E5”. So E5 is not large proteins or small proteins, it’s not non-coding RNAs — according to this document, E5 is made of exosomes and exosomes alone. Independently, a group from the Smidt Heart Institute published in July their finding that exosomes from embryonic plasma had a profound rejuvenating effect, while plasma without exosomes had no effect. (Akshay alerted me to that paper when it came out.)

This new preprint reports on experiments using E5 from piglets to rejuvenate a total of 14 rats. Epigenetic markers, cognitive, physical, and metabolic markers all improved impressively. Some of the limitations of Katcher’s experiments remain to be resolved. Because of small numbers and killing of test rats to obtain tissue samples, we still do not have a strong indication to what extent the restoration of all these markers of age translates into increased longevity.

A major finding is that exosomes from a young pig can rejuvenate an old rat. It is remarkable, very lucky, and certainly not to be taken for granted that such specific signals can cross species lines. What a horrible situation we would have been in if it turned out that the best anti-aging treatment available depended on harvesting the blood of babies.


There were only 14 treated rats in this round of experimentation. 6 males in round 1 were sacrificed so that methylation of their organs could be tested. 6 female and 2 male rats in round 2 were kept alive long enough for their blood to be drawn for methylation testing, but not long enough to determine their lifespans.

Methylation clocks: For 6 treated rats, age 109 weeks, these were the average epigenetic ages by tissue:

Tissue Epigenetic age
liver 28 weeks
blood 39
heart 58
hypothalamus 82

The preprint does not provide error bars for these average epigenetic ages, but I expect (based on the published minimum and maximum, and the small sample size) that the 95% confidence intervals are broad, corresponding to rejuvenation by anything between 30% and 70%.

The fact that various tissues are rejuvenated by exosomes indicates that exosomes are a medium for transmission of age information through the body, and that young exosomes in the blood are able to induce system-wide changes in gene expression.

It’s common for medical gerontologists to frame their work in terms of their favorite evolutionary theory of aging, which is always the 70-year-old, conservative theory that won’t raise any hackles. I’m grateful that Dr Katcher is unabashed about framing his medical research in evolutionary terms, and that his evolutionary theory agrees with mine. Aging is a timed program of self-destruction, coordinated through the body by chemical signals, and evolved because ecological communities are more stable when individuals have predictable lifespans.

The hypothalamus is a neuroendocrine region of the the brain that has been associated with circadian timekeeping (day-night) and also with tracking biological age. The fact that gene expression in the hypothalamus was rejuvenated to a far lesser extent than the other three tissues suggests that hypothalamus plays a special time-keeping role, and that perhaps the hypothalamus is a source for these exosomes in vivo. It may be that the hypothalamus gradually erases half the epigenetic gains from the E5 treatment over time. From this and previous tests, we still have limited data on long-term benefits of E5 treatment.

Blood markers:

“we measured the levels of the following biomarkers on 30, 60, 90, 120 and 155 days from the start of the experiment: bilirubin, serum glutamic-pyruvic transaminase (SGPT) and serum glutamic-oxaloacetic transaminase (SGOT) to monitor liver function; triglycerides (TG), HDL and cholesterol to monitor risk of atherosclerosis and heart disease, and liver function as well; glucose to monitor the pancreas and diabetes; and creatinine and blood urea nitrogen for kidney function. The levels of all these biomarkers in the treated old rats were altered towards the values of young rats, without exception.”

(Note: Blood sugar and triglycerides in treated old rats was reduced to levels indistinguishable from untreated young rats. In humans, Type 2 Diabetes is a primary means by which the body destroys itself. — JJM)

Cognitive recovery: Treated rats learned to escape from a Barnes maze faster than old rats.

Oxidative stress: ROS markers were restored close to a young state. Antioxidant enzymes GSH, SOD, and catalase were increased close to youthful levels.

Inflammation: TNF-α and IL-6 were reduced close to the levels of young rats. Nrf-2 was increased close to the level of young rats.

Sarcopenia and muscular fitness: Grip strength of treated rats rose within a week to levels comparable to young controls, and remained strong for at least 30 days.

A new/old biological clock based on glycation of antibodies

IgG is a blood component, another name for “antibodies”. These bind specifically to viruses and bacteria and tag the cells for attack by phagocytes, another immune component.  Antibodies are not cells or globules but individual Y-shaped protein molecules with both arms of the Y able to bind to the (same) antigen, and the stem of the Y modified to attract a phagocyte only when the other end signals that it’s “got one”.

Immune aging is a big deal, as it was identified as a prime cause of aging by Roy Walford already in the 1960s. Greg Fahy pushed the theory into a therapeutic concept four years ago with his TRIIM program. With TRIIM, the idea was validated that a rejuvenated immune system can signal rejuvenation to epigenetic age and, presumably, other aspects of aging.

Dr Fahy also subscribes to a programmed theory of aging, and he personally offered me my first opportunity for a wide audience in the anti-aging community in 2010.   Mitteldorf_Evolutionary-Origins-of-Aging.pdf

Hence, “immune age” is a promising target for measuring success of anti-aging interventions, independent of methylation age. This idea has been realized with a Glycan-Age clock. Antibodies in the blood are coated in complex carbohydrate (sugar) molecules, or glycans, and the structure of the glycans carry information, analogously to the way that DNA or proteins carry information via their sequencing. The particular sugars attached to antibodies change with age, and they become important signals promoting inflammation later in life. Clocks based on glycation of antibodies are as old as methylation clocks, though they have been developed to a lesser extent.

It is surprising, perhaps, that glycan age was never measured in the TRIIM study, but Katcher’s group added glycan age to the tests on his treated rats.

“a significant (p < 0.05) reduction in the relative abundance of the pro-inflammatory agalactosylated IgG2a glycoform (G0) was recorded. This was accompanied by a simultaneous upsurge in the antiinflammatory digalactosylated glycoform (G2)”

Is there such a thing as “biological age”?

The preprint includes the statement,

“However, in the context of aging and rejuvenation, it is crucial to differentiate between improved health or organ function, which could be achieved via medication or surgery, and genuine molecular age reversal.”

Intuitively, we all feel that this must be a legitimate concept. In our experience, we can look at most people and form a pretty good idea how old they are, and this translates into reliable expectations about their stamina, resistance to disease, and ability to cope with a fundamentally new environment.

But this is because we’ve mostly encountered natural humans. We know, for example, that movie stars can have expensive surgeries and skin treatments that make them look decades younger. We don’t expect that people who have had a face lift will live longer or have more youthful vitality.

It is perfectly possible to imagine anti-aging treatments that rejuvenate the liver but not the nervous system, or that slash the risk of heart disease but increase the risk of cancer. Hence, it is advisable to report a variety of functional and metabolic tests rather than relying on any one measure of “biological age”.

What is measured by the methylation clocks?

For more than a decade, I have been committed to the paradigm that the body’s gene expression changes with age, and that changing gene expression is a driver of aging. Methylation, as a convenient surrogate for gene expression, would then seem to be a reliable measure of biological age.

But in the last four years, I’ve entertained questions about this paradigm, and I am no longer absolutely committed. The crucial question is: Does gene expression change as a driver of aging, the body destroying itself with inflammation and by scaling back autophagy and repair? Or does gene expression change in response to damage accumulated with age, scaling up repair functions to mitigate damage?

As recently as four years ago, I wrote confidently that the former greatly predominated. Epigenetic changes drive aging. On this basis, I was confident that if you change the methylation clock, you must be changing the driver of aging, and this was bound to lead to longer lifespan.

I have since become convinced that both kinds of changes in gene expression accrue with age, and that there is no easy way to tell the drivers from the responses. I first blogged about this in response to the GrimAge clock. GrimAge is the best predictor we have of remaining life expectancy, but it is based not just on drivers of aging but also on the body’s distressed response to toxins and cell-level damage.

The bottom line is that epigenetic clocks are the best measure of life expectancy that we have, and that they are improving every year, but while uncertainty remains, it is well to supplement methylation clocks with specific tests of metabolic function, stamina, inflammation, and cognitive performance.

Six new methylation clocks

Prof Horvath (presumably it was Horvath) trained four new clocks on rat tissues plus two new clocks that were cross-trained on humans and rats. Since rats have a maximum lifespan of 3.8 years and humans 122.5 years, one logical way to train a clock that works for both species would be to scale the rat ages and human ages so that one rat year is about 32 human years. This is just what Horvath did.

But in addition, he trained a second combination clock (rat and human) to read in absolute years. This presumes there are changes in epigenetics of a very old rat (say, 3 years) that are parallel to epigenetic changes taking place in a very young human (also 3 years). A strange concept, indeed! What kinds of changes would we expect to take place in a 3-year-old rat and simultaneously in a 3-year-young human? My guess is that the training program is selecting those sites that behave very differently in rat and human, and that results derived for rats with this clock will not translate.

But I would suggest that even the time-scaled clock (based on 32 years of human life = 1 year of rat life) is suboptimal. This is because different events in development and aging may occur at different points in the life cycle. A male rat reaches sexual maturity at about 8 weeks. Using a 32x multiplier, this would correspond to a 5-year-old boy. But in fact, boys don’t reach sexual maturity until well into their teens.

Prof Horvath has been developing methylation clocks for more than a decade, and his work has had a deep impact on the science of aging. To my knowledge, all his clocks are based on straight-line changes in methylation over a lifetime, and I have counseled that this is an unnecessary limitation on the clocks’ accuracy. More realistic clocks can be constructed based on parabolic curves, or even the simple expedient of grafting two straight lines together to create a “spline”. He has been creating cross-species clocks for the last few years, and to my knowledge these have all been based on a single scaling factor for each species. My recommendation is that here it is even more important to project the life cycle of one species onto another using a flexible mapping, pegged at several intermediate time points which are determined in a mathematical optimization process. Cross-species clocks will be of increasing importance as we translate mouse studies to develop experimental protocols for human medicine.

Unanswered questions

A few paragraphs up, I referred to a major issue in methylation clocks: A given intervention affects a set of methylation sites that are all associated with younger age. Which of these are actually beneficial, leading to life extension; and which represent reversion of the body to a younger state when it had less damage to deal with? For the latter, the younger state corresponds to repair mechanisms that are less active (presumably because they are less needed), and “rejuvenation” to this state could actually lead to shorter life expectancy.

To date, the best methylation clocks have been calibrated by how well they predict future life expectancy. This is, of course, very useful information, but it is all based on naturally-aged individuals. Our interventions that reduce methylation age may not increase lifespan.

In two previous rounds of Katcher’s experiments, males experienced more epigenetic rejuvenation, but they were killed before their lifespans could be measured; while females experienced less epigenetic rejuvenation, and their lifespans were measured to be only modestly extended relative to controls. Unexplained is the fact that the controls lived far longer than Sprague-Dawley rats are usually expected to live, and that a single treated rat lived far longer than 7 others that received the same treatment. These are all curious observations that warrant further investigation.

The failure of Katcher’s rats last year to live as long as their methylation age predicted should be a motivator for future research, studying not just the interventions but also the clocks.

We can try to classify different methylation sites based on theory; but if we want to address the question head-on, we need different experimental interventions that affect different CpG sites. Then we can see which CpG sites, when artificially methylated or demethylated, lead to longer lifespan. Hence, Katcher’s experiments provide a unique opportunity to determine in practice: which methylation sites, when reverted to a younger age state, actually extend lifespan? This is an experimental solution to the stickiest problem in aging clocks: which epigenetic changes correspond to drivers of aging and which correspond to responses to damage?

Do we have methylation for Sima, Katcher’s record-setting long-lived rat? If so, we might compare methylation profiles for Sima to his cage-mates that were treated the same way, yet died many months earlier. This is a start, but what we really need is an experiment in which about a hundred rats (M and F) are treated with different E5 protocols, their methylation profiles are monitored over time, and they are also permitted to live until their natural end. Manipulated methylation profiles can then be correlated with life expectancy, and from this a uniquely useful methylation clock can be constructed.

For me, the bottom line is the same thing I’ve been saying since the beginning: Katcher’s research is the most promising line of investigation in the field of anti-aging medicine, and it is a travesty that it is taking place under a single investigator’s direction, and with very limited resources. It will take extensive lab resources to optimize E5 dosage and timing, and to try combining addition of young exosomes with removal of old exosomes. This same research can be the basis for the first methylation clock that is known to predict lifespan for rejuvenated animals.

I would like to see the abundant money that is now in anti-aging research re-directed into rejuvenation via exosomes, and I would like to see Katcher sharing his techniques openly with the research community.

Exosomes and their Potential for Rejuvenation

Extracellular Vesicles (EVs) have only been studied in the 21st century. Think of them as natural lipid nanoparticles, or endogenous viruses. They transmit information around the body and they’re small enough to be exhaled and carried in the air to communicate with other individuals and even other species. EVs are encapsulated in fats that facilitate entry into cells, and inside they contain proteins, RNA, DNA, lipids — all of which carry information. EVs are a universal biological language, a barely-explored medium of communication.

Last week, an extraordinary paper was published demonstrating the rejuvenation potential of EVs from very young animals injected into old animals. The authors come from Smidt Heart Institute in Los Angeles, and their primary field is cardiology, not gerontology, so they focus on heart-derived EVs and benefits for rejuvenating the heart. But the benefits they observe go well beyond the heart and include lifespan in rodents and rejuvenation of human cells from rat-derived EVs.

This research should be understood in the light of parabiosis experiments by the ConboysRando, and others and in particular to the rejuvenation technology of Harold Katcher. I am grateful to Harold’s partner, Akshay Sanghavi, for alerting me to this publication. Katcher’s blood-derived rejuvenation serum is called E5, and Akshay has told me that E5 includes EVs as well as a wide range of proteins. The current article suggests the possibility that it is wholly the EVs that are responsible for the benefits of E5. Indeed, the authors repeated their experiments with (1) whole blood plasma, (2) plasma with the EVs removed, and (3) just EVs. They found rejuvenation effects associated with whole plasma and with EVs alone, but not with plasma minus EVs. If this holds up, it is the study’s most important finding. Akshay told me that EVs are easier and cheaper to isolate than the molecular constituents of E5 (proteins).

For all those involved in parabiosis and plasma exchange research, I suggest that it should be an immediate priority to replicate the Smidt findings that all of the rejuvenating power of young blood is contained in EVs. The Conboys might be interested in asking whether the pro-aging effect of plasma infusions from old to young animals is also an effect of EVs.

The state-of-the-art EV separation technique for large throughput is called an acoustic nanofilter. Ultrasound pressure can be tuned to separate a particular size of particle in a specially-designed medium. If I were advising Yuvan, I would suggest that they develop an expertise in acoustic nanofiltration ASAP as a next generation replacement for their plasma fractioning technology.


Exosomes are the most common type of EV, and probably the most relevant to aging applications. They seem to be a general vehicle for inter-cellular and inter-individual communication. The study of exosomes is in its infancy, but it is already known that exosomes are tagged in a way that recipient cells can distinguish and choose which exosomes to pick up and “read the message”.

One area of exosome activity that has been studied is the communication of antigens to the immune system. A particular type of exosome includes foreign proteins that are potential invaders, and they are shared not just within the body but through the air. When you think of herd immunity, consider not just the contagion of people who carry disease but also the information about what diseases are in the air that is transmitted in social interactions.

When communities and whole countries were locked down during COVID, one consequence was to prevent uninfected individuals from learning and preparing an immune response to the virus through the sharing of exosomes.

“In the nervous system, exosomes have been found to help in myelin formation, neurite growth, and neuronal survival, thus playing a role in tissue repair and regeneration.” [refrefrefrefrefref] “It has been demonstrated that the mesenchymal stem cell exosomes themselves can act as a therapeutic agent to help [repair] tissue injury.” [refrefrefrefrefref]

“Microvesicles” are, as far as I can tell, a kind of fat exosome. They are as large as bacteria, whereas exosomes are closer to the size of viruses, but the range of contents and hypothesized functions is the same.

Apoptotic Bodies — another potential target for EV therapies

There is also a specific EV type that apoptotic cells send to induce apoptosis in other cells. Apoptosis is programmed cell death, playing an important role in self-destruction of diseased and cancerous cells. However, aging involves a cascade of apoptosis in healthy cells that leads to loss of muscle tissue (sarcopenia) and brain cells (neurodegeneration). Apoptotic bodies are the type of EV that triggers the apoptosis cascade, and I speculate that removing apoptotic bodies from the blood will have future potential as an anti-aging therapy. Apopototic bodies are even larger than “jumbo” microvesicles, so it should not be difficult to separate them in medical applications. It is a reasonable guess that the effectiveness of senolytic therapies is directly related to a reduction in apoptotic bodies, which are secreted by senescent cells. (These are my own speculations, not mentioned in the paper that I’m reviewing here.)

The new research

Rejuvenating effects of young extracellular vesicles in aged rats and in cellular models of human senescence by Grigorian-Shamagian et al, in Nature Scientific Reports. Senior author = Eduardo Marbán.

The authors extracted EVs from neonatal rat hearts, and applied them to cell cultures and injected them into older live rats (22 months). They used whole plasma and plasma minus EVs as controls, establishing that it is the EVs that carry the benefits. They extracted corresponding EVs from heart cells of neonatal humans, and applied them to cultures of human cells.

They found EKG evidence that hearts were functioning better in treated animals. Insulin resistance was blood sugar were reduced.

Exercise endurance increased in treated animals, while declining progressively in controls.

They found that pliability of tissue was restored toward young levels. Hearts, lungs, muscles, and kidneys all improved their function in treated rats.

Tissue samples (muscle and heart) look better in ways that I don’t pretend to understand.

Old rats were treated with four monthly infusions, then kept for 16 weeks more before sacrificing all animals. During those 16 weeks, 6 of 14 control rats died and only 3 of 36 treated rats. My calculation (Fisher’s exact test) indicates life extension with a 99% confidence level.

Tests were conducted on two human cell cultures: fibroblasts from middle-aged donors treated with human EVs, and cardiac cells treated with rat EVs. Fibroblasts increased reproductive capacity, decreased apoptosis, and a more youthful transcriptome. Similarly, cardiac cells exhibited better self-assembly and a more youthful transcriptome.

The bottom line

In case you haven’t yet read between the lines, I’m excited about this work. I would have liked to see results of some Horvath clocks and a hat tip to the Katcher preprint, but in other respects, I find the research quite thorough and convincing. A major question in plasma exchange research has been the identification of the active component, out of thousands of protein species. This is only one study, but it suggests that researchers look at the activity of EVs rather than proteins or RNAs.


A crucial question is whether EVs from other mammals can be used to rejuvenate human tissues. This study suggests yes, but the demonstration was limited to cell cultures and no direct comparison was made between treatment with human-derived and rat-derived EVs. The author has not responded to my email requesting any data relevant to this question. If animal EVs work in humans, the therapeutic market may soon open; EVs should be easy to extract from the blood of young animals that are being slaughtered for meat. If not, we face a major ethical dilemma, as aborted and stillborn infants will support only a tiny fraction of the potential demand for rejuvenating EVs. EVs from young human donors can probably be extracted but it’s a lot to ask of our children. The authors suggest that EVs could be manufactured from human cell cultures. “Given that allogeneic CDCs [cardiosphere-derived cells] are already in advanced testing and have proven safe to date, such cells can be used as manufacturing platforms for EVs, enabling rapid progress to clinical testing in a variety of aging-related disorders.”

Cellular Rejuvenation Without Yamanaka

The Yamanaka factors (four proteins, abbreviated OSKM) are a treatment that can completely reset any ordinary, functional cell into a pluripotent stem cell, able to regenerate tissues of any type. In the process, the cell loses epigenetic markers of age and reverts to an embryonic state. But differentiated cells are differentiated so they can be functional. Taking them all the way back to stem cells destroys their function. Remarkably, differentiated cells can be taken part-way back, becoming younger but not losing its functional identity. This has been done with pulsed doses of OSKD or with just three of the four factors, OSK. Last year, the Harvard laboratory of David Sinclair reported success rejuvenating the aged retinas of mice with this process.

Rejuvenating a whole body in this way is impractical. Precisely controlling the dosage and the timing of OSK to each cell cannot be achieved outside a Petri dish. To make matters worse, some of the treated cells have a tendency to grow into tumors. 

Sinclair has been searching for other chemicals (besides the Yamanaka factors) that might have a rejuvenating effect at the cellular level without risking cancer and without problematic sensitivity to dosage and timing. With modern, computerized laboratories, it is practical to test thousands of chemicals shotgun-style in separate cell cultures, automating both the delivery of chemicals and the measurement of outcomes.

Recently, Sinclair reported the results of this process: Of the many combinations assayed in his shotgun experiments, six were identified that successfully rejuvenated cells in culture without the cells losing their identities and without causing cancer. “Thus, rejuvenation by age reversal can be achieved, not only by genetic, but also chemical means.”

This sounds like breakthrough science. Let’s look more closely.

What is the evidence that the cells are rejuvenated?

The new Sinclair article cites two lines of evidence

  • Nucleocytoplasmic compartmentalization (NCC)
  • An aging clock based on which genes are being transcribed


NCC was a new idea for me. There are certain proteins that belong in the nucleus, and they tend to leak into the surrounding cytoplasm as one sign of aging in the cel. 

Aging clocks based on transcription are a promising idea but not as well developed as methylation clocks.

Maybe I’m too suspicious, but I’m inclined to ask: Why didn’t Sinclair choose one of the well-established, validated aging clocks as a target, for example the robust Horvath multi-tissue or multi-species clocks? If he was going to use a transcriptomic clock, why did he have to invent a new one, rather than use Lehalier’s clock? If he was going to introduce a new clock, why wouldn’t he validate it by reference to established aging clocks?

I wonder if there is an untold history of this research, in which the Sinclair lab tried and failed to rejuvenate cells as measured by the more conventional aging clocks. 

Aging is not cell-autonomous

In different laboratories, aging research has been pursued via two pathways: top-down and bottom-up. Top-down researchers are looking for central coordination of the age state of the body, broadcast probably via signal molecules in the blood. (Michael Levin thinks that it may also occur through persistent patterns of electrical voltage.) Bottom-up researchers are looking for ways that individual cells lose focus and function with age.

The deep problem with Sinclair’s approach is that it derives from an overly reductionist perspective. Aging of the body is reduced to aging of individual cells. Aging of cells is understood as a blurring in the precision chemistry that the cell relies on for optimum health. 

Other researchers (including the Conboys, Harold Katcher, and Izpisua Belmonte) are pursuing the top-down approach to rejuvenation, based on a paradigm in which aging is coordinated systemically, and not controlled at the cellular level. Interventions are in the bloodstream, through which biochemical signals transmit information about the age state of the body

Sinclair repeats without qualification his prejudice that “aging is a loss of epigenetic information.” He cites his previous experiments as support for “the idea that a loss of epigenetic information, resulting in changes in gene expression, leads to the loss of cellular identity.” But we have learned that methylation changes and other epigenetic changes with age are a redirection as much as a loss of information. Some of the changes are predictable and programmed, and, indeed, the idea of a methylation clock depends on the regularity of methylation changes

For the future

I like Sinclair’s idea for high-throughput screening of rejuvenation molecules. I would add that combinations should be tried after the first step of identifying promising single candidates. The example of OSKM demonstrates that combinations are able to achieve what no single chemical can do alone. In the discussion section, Sinclair indicates that his lab is already doing this, and that the work is guided by an investigation of the mechanisms of action of the chemicals that his screening procedure discovered. 

The present work was done with human fibroblasts (cells from connective tissue) and the discussion section mentions that ongoing work to repeat the test with cells from other tissues. The screening procedure requires the extensive laboratory resources that Sinclair is able to command. 

A consistent theme in my work is that aging is evolved, programmed, and centrally coordinated. I’ve cited diverse evidence for this, based on observation, not theory. As a paradigm to guide anti-aging research, there is an additional reason to research central control rather than cell-autonomous interventions. That is that engineering the rejuvenation of every cell in the body is a daunting task. In contrast, if aging is centrally controlled and communicated via biochemical signaling, we can expect that modification of the signaling environment ought to be far more accessible goal.

Mortality Spiked in the Pandemic Years, over and above COVID deaths

There is still confusion about what happened during the Pandemic. The public health response was innovative, defying historic precedent. Was it effective? One reason it is hard to know is that the health records are administered by the same agency (CDC) that managed the pandemic. Death certificates are subjective, and the listed cause of death is prone to error and outright manipulation. Death counts are often reliable, even when cause of death is debatable.

The CDC has recently released death counts from 2021, and they are even higher than 2020. In addition to the deaths attributed to COVID, there were 155,000 excess deaths in 2020-21 among younger Americans. It is a reasonable conjecture that these deaths were related to the circumstances of the pandemic. Below, I lay out the numbers and explore some interpretations.

The CDC keeps track of how many people die each year, their ages, the place, and the cause of death. Data is released to the public 18 months after year end through a searchable resource at https://wonder.cdc.gov/. All-cause mortality data for 2021 was recently released, Unexplained high death rates in 2020 continued and accelerated in 2021. On top of the death burden of COVID, there were 155,000 deaths in people under 65. (I focus on people age 15-64, because COVID deaths are relatively rare. Only about one fifth of COVID deaths were under 65.) 

Deaths in this younger age group were up 19% in 2020 compared to the average of the previous five years. In 2021, deaths were even higher, up 30% compared to the same baseline. Most of this is not accounted by COVID. Excluding COVID deaths, there is an excess of 17% in 2020 and 23% in 2021. That’s 155,000 unexplained deaths. COVID deaths in this age group were 220,000 over two years, so the unexplained deaths are two thirds of the official COVID death count. Excluding the oldest group (55-64 years), the unexplained deaths are actually greater than the COVID deaths. 

Word on the Hill is that official counts of COVID deaths were overstated, so the difference becomes more dramatic and the cause more mysterious. These numbers are far outside year-to-year fluctuations (~20,000) or statistical uncertainties. 155,000 Americans is almost three Vietnam Wars. The last time younger Americans were dying at the 2021-22 rate was World War II.

Proportionally, the excess deaths were concentrated in the 25-44 age group. But the highest absolute numbers were in the oldest age group (55-64), as is almost always the case. Excess deaths among males and females were skewed 64:36, but this is actually normal; in this age range, men are dying at a much higher rate than women. 

The biggest wave of excess deaths occurred in August through October of 2021. This was not a period of heavy vaccination and it is off-season for respiratory infections. The second biggest wave was December, 2020 through February, 2021. This was a time when COVID was reaching its second wave, but remember that COVID deaths are excluded in this tabulation. A speculation is that the initial release of the Pfizer vaccine was so toxic that the formula was changed within a few weeks. This hypothesis is supported by VAERS data, which were surging during these same months. 

What caused 155,000 excess deaths?

This count is in addition to deaths coded COVID-19. The increase began before the vaccine rollout, so at least the 2020 portion cannot be blamed on vaccine deaths. An obvious place to look is in unintended consequences of the closures and lockdowns ordered by Federal health officials as an emergency response.

CDC categorizes deaths by cause, (but then doesn’t break this down by age or month). The top ten causes of death from 2019-21 were, in order,

  1. Heart disease
  2. Cancer
  3. COVID-19
  4. Accidents
  5. Strokes
  6. Lung diseases including emphysema and COPD 
  7. Alzheimer’s disease
  8. Diabetes
  9. Chronic liver disease and cirrhosis
  10. Kidney failure

Some of these categories increased significantly in 2020 and 2021 compared to 2019 as a baseline. There were 75,000 excess heart attack deaths over the two years. Some of this may be attributed to heart attack patients forgoing emergency care because they feared COVID in the hospitals. Deaths classed as “accident” were up by 28,000 in 2020 and a striking 52,000 in 2021. Deaths listed as “suicide” seem to be intentionally undercounted, but (surprisingly) they did not increase in 2020 or 2021.

Drug overdose deaths might be expected to increase during lockdowns. Drug overdose deaths are not listed as a separate category by CDC, but from this CDC announcement it is possible to estimate 15,000 extra deaths per year in 2020 and 2021. According to CDC policy, some of these may have been listed as “accidents”. 

2020 XS 2021 XS
1. Heart disease 37,921 36,506
2. Cancer
3. COVID-19
4. Accidents 27,915 51,895
5. Strokes 10,259 12,885
6. Lung diseases
7. Alzheimer disease 12,743
8. Diabetes 14,541 15,647
9. Liver disease 7,284 12,227
10. Kidney failure
Total accounted for 110,663 129,160

Excess deaths in the above table are differences compared to a 2019 baseline. Empty cells connote counts that were negative or not significantly positive.


A classical public health response to a pandemic was advocated by Dr David Katz, by Dr John Ioannidis, by Dr Harvey Risch, by the Great Barrington coalition, and several other prominent epidemiologists in the spring of 2020. Their proposal would have been to isolate and protect the most vulnerable, while allowing the virus to spread among young, healthy people who would almost certainly survive the disease, and would earn the herd immunity needed to end spread of the virus. 

Instead, US public health agencies tried a radically new approach, constricting commerce, shuttering churches, concert halls, restaurants, and cultural institutions, encouraging millions of people to stay in their homes, masking the healthy as well as the sick, discouraging experimentation with early treatments that seemed promising while rushing vaccines through testing into mass distribution.

We cannot know what events might have unfolded had these mainstream public health experts prevailed, but the results do not reflect well on choices made by our CDC and NIH and the White House task force. COVID deaths were high in 2020 despite the restrictions, and twice as high in 2021 despite the vaccines. Consequences of the lockdowns and closures were not limited to the huge economic loss ($14 trillion according to this USC study), but included indirect deaths that were higher in young people than deaths from COVID itself.

Let us encourage our public health officials to learn from these mistakes.

After writing this, I was referred by a reader to the work of Denis Rancourt, a Canadian researcher. In this video he does a much more thorough job that I have done, starting with US all-cause mortality data but digging deeper by locale and interpreting through a perspective he has gained from international data.

Data tables

Z scores are a statistical measure of how improbable each month’s number is based on the distribution of death counts in the baseline years 2015-2019. Z=2 is the classic 95% confidence test. Z=5 means a 1 in a million fluke — in other words, this was certainly not due to chance.

“Z scores” is defined in the caption above.

These are monthly death counts over and above the corresponding monthly average from 2015-19

These are monthly death counts over and above the corresponding monthly average from 2015-19