I Scored 99 Symptoms Before I Started. Forty-Six Days Later I Scored Them Again.

622 points to 268. What moved, what didn’t, and the actual biology of why a change of fuel can quiet a body — explained from the beginning, with the research attached.

Start here: the eight words you need

I am going to define every term the first time I use it, and I will define them again when they reappear. You should never have to stop reading to look something up.

Ketosis. The metabolic state in which your body runs mainly on fat-derived fuel instead of sugar. It is not a disease and not a trick; it is one of two normal modes the human body has always had. Analogy: a hybrid car that can run on gasoline or electricity. Most modern people never leave gasoline mode.

Ketones. The small fuel molecules your liver manufactures out of fat when carbohydrate is scarce. They travel in the blood and cross into the brain, which cannot burn fat directly but burns ketones happily.

Beta-hydroxybutyrate (BHB). The main ketone, and the one you can measure. A drop of blood on a test strip gives you a number in millimoles per liter (mmol/L). Most people eating normally sit near 0.1. A typical “keto diet” produces somewhere around 0.5 to 1.5. The therapeutic range — the level this experiment is about — is above 4.0, which is difficult to reach and considerably more difficult to hold. Every time you see BHB in this article, it means that number.

Mitochondria. The power plants inside your cells — typically hundreds to thousands per cell, and most numerous in the tissues that never rest, like heart and brain. They take fuel and oxygen and produce ATP, the actual currency of cellular work. Analogy: if the cell is a factory, mitochondria are the generators in the basement. When they run poorly, every department upstairs gets dimmer, and the departments with the biggest electricity bills — your brain, your heart — notice first.

Mitochondrial biogenesis. The process of building new mitochondria — adding generators to the basement rather than tuning the existing ones. This is slow, structural work, and that distinction matters enormously later in this article.

Inflammation. Your immune system’s repair-and-defense response. Useful in bursts — it is how a cut heals and an infection is cleared. Destructive when it never switches off, which is the state most modern chronic disease shares.

The NLRP3 inflammasome. The hardest term here, and the most important, so take it slowly. An inflammasome is a molecular alarm assembly inside your immune cells. NLRP3 is the best-studied one. When it detects certain danger signals — uric acid crystals, certain fats, cellular debris — its parts snap together like a cocked trigger and release inflammatory messengers called cytokines (chemical shouts that tell the rest of the immune system to escalate). The best-known cytokine it produces is interleukin-1 beta, written IL-1β. Analogy: NLRP3 is a smoke alarm wired to the whole building. In healthy function it goes off for a real fire. In chronic inflammatory disease it is a smoke alarm stuck screaming with no fire in the building — and every occupant, meaning every tissue, lives in that noise. Aching joints, inflamed tendons, headaches, swollen airway tissue: that is what a stuck alarm feels like from inside.

HDACs (histone deacetylases). Enzymes that control which of your genes are readable. Your DNA is spooled around proteins called histones; how tightly it is wound determines which instructions can be read out. HDACs tighten the spool, silencing genes. Analogy: a library where HDACs are the librarian locking certain books shut. Inhibit the librarian, and certain books become readable again — including, as you will see, the book for building brain-repair proteins.

BDNF (brain-derived neurotrophic factor). A protein that supports the survival of existing neurons and the growth of new connections between them. Informally: fertilizer for the brain. Exercise raises it; so, it turns out, does BHB.

Why I scored myself before I knew the answer

A before-and-after is worth almost nothing if the “before” is written afterward. Human memory is not a neutral archive; it edits toward whatever story you are currently telling. If I had improved for six weeks and only then tried to describe how bad things had been, I would have produced a testimonial, and you would have been right to discount it.

So on the morning of June 17th — before a single day of the attempt — I listed every complaint I had and scored each one from 0 to 10, where 0 means gone and 10 means the worst it has ever been. Ninety-nine items, grouped into thirteen sections: skin and hair, body composition, joints, headaches, breathing, eyes, smell and taste, hearing, cardiovascular measurements, urinary, cognition, mood and personality, sleep, and energy.

I want to be honest about the state of mind that list came from. I am sixty. Almost none of these problems announced themselves. They arrived by degrees across years, which is precisely why I had stopped calling them problems and started calling them normal — and the person changing is always the last to notice. I have watched exactly this in patients for thirty years: they report “nothing much,” and then you ask specific questions and a catalogue falls out. It took writing my own catalogue to see myself the way I would have seen a patient.

The June total: 622.

On Day 46 — August 1st — I re-scored every item on the same scale. One rule for myself: where I hesitated between two numbers, I took the worse one. A reader should assume I am biased toward finding improvement, so I scored against myself deliberately. The new total: 268. A 57% reduction in total symptom burden in forty-six days.

The results, section by section

Section totals flatten things, so here are individual rows, quoted from the scorecard as written.

“Difficulty bending forward to tie shoes” — scored 10 in June. A physician, sixty years old, who could not comfortably tie his own shoes. Day 46: 0. “Difficulty putting on socks”: 10 to 0. “Neck pain, rotation limited to about 45 degrees, disturbs sleep”: 10 to 0. “Tendonitis at inner elbows”: 10 to 0. “Exertion-related pain flares — hands, wrists, elbows, for days”: 10 to 1. “Exercise-induced pain that forces you to stop”: 10 to 0. “Pain when standing after sitting more than 30 minutes”: 10 to 2.

“Post-occipital headaches, base of skull, constant”: 8 to 0. “Severe snoring”: 7 to 0. “Noisy breathing at rest”: 7 to 0. “Difficulty on inspiration, feels constricted”: 7 to 0. Smell and taste, both scored 5: both 0.

And the ones that are harder to publish. “Mental fog — dull, foggy, not sharp”: 9 to 2. “Very poor concentration”: 9 to 3. “Very irritable”: 8 to 2. “Anhedonia / emotional blunting — numb, things don’t bring joy”: 8 to 3. And a row I included because a scorecard that omits the embarrassing entries is not a scorecard: “Yelling at the cats” — 9 in June, 2 now.

Three rows are instruments rather than opinions. My fasting fingerstick glucose went from 130 mg/dL to 89. (Fingersticks, not laboratory draws — no physician diagnoses anything from single readings, and I never ordered confirming labs. But I have read these numbers for thirty years, and 130 fasting is not a number you shrug at.) Heart rate variability, a rough index of autonomic recovery, went from 54 to 99 milliseconds. And the drop of blood itself changed: in June I wrote one word in the notes column — THICK — because it beaded slowly and dark. On Day 46 I wrote: thin.

Why would a change of fuel do this? Four mechanisms, ranked by how sure I am

This is the section I most wanted to write, because “I did keto and felt better” is worth nothing without a plausible mechanism — and because the mechanisms here are not folklore. They are published, replicated, and specific. I have ranked them from most to least certain, and I have tried to be as clear about the limits as about the findings.

1. BHB is a better fuel for a struggling brain (certain, and fast)

The least controversial mechanism. Beta-hydroxybutyrate — BHB, the ketone you measure in blood — is not a marginal backup fuel; it is a preferred one for the brain under many conditions, and it yields energy efficiently. When glucose delivery is unstable — which is what insulin resistance produces as it develops over years — the brain experiences intermittent fuel shortage. Ketones bypass that problem entirely, because their uptake does not depend on the same insulin-driven machinery.

This is almost certainly what explains the fastest changes in my scorecard: fog, focus, concentration, energy. Those are the rows that move in days, not weeks, and they move in the direction of the number. I have watched this repeatedly in my own diary — at BHB 2.5 I am foggy and sluggish; above 4 the fog lifts, and above 5 I am sharper than my ordinary baseline. This is not repair. It is refueling. It should be fast, and it is.

2. BHB switches off a specific inflammation alarm (established, and fast)

This is the mechanism I believe explains most of the joint, headache, and breathing rows — the largest movements on the whole scorecard — and it is the one with the most direct experimental support.

Recall the NLRP3 inflammasome: the molecular alarm assembly inside immune cells that, when triggered, releases inflammatory messengers — cytokines — chief among them interleukin-1 beta (IL-1β). In 2015, a team publishing in Nature Medicine showed that BHB directly blocks NLRP3 activation [1]. The specificity of that finding is what makes it persuasive. BHB did it; acetoacetate — the other main ketone, structurally similar — did not. Butyrate and acetate, related short-chain fats, did not. And it did not depend on the usual starvation-response suspects: not AMPK, not reactive oxygen species, not autophagy, not glycolytic slowdown. BHB appeared to act on the assembly itself — preventing potassium efflux from the cell and stopping the alarm’s components from clustering into their active form.

In other words: this is not “ketones are healthy, and inflammation went down.” It is a molecule fitting a specific lock.

Two years later, the same line of work went after a disease that is essentially pure NLRP3 in action. Gout is caused by uric acid crystals triggering that alarm inside neutrophils — the immune cells that produce the flare — and the result is joint destruction and extraordinary pain. In Cell Reports in 2017, researchers showed that a ketogenic diet raising BHB reduced urate-crystal-induced gouty inflammation, that BHB inhibited NLRP3-dependent IL-1β release from neutrophils, and — importantly — that it did so without crippling the animals’ ability to fight bacterial infection [2]. That last detail matters more than it might seem: it distinguishes a targeted quieting of one alarm from general immune suppression, which is what most anti-inflammatory drugs do and why they carry the risks they carry.

Now look back at my scorecard with that in mind. The seventeen-item joints section fell 154 to 31 — tendonitis, enthesitis, the pain after sitting, the flares lasting days. Constant headaches at the base of the skull: gone. Snoring, noisy breathing, the constricted feeling on inspiration — all of which are, mechanically, inflamed and swollen airway tissue: 42 to 3. These are precisely the tissues where a stuck NLRP3 alarm expresses itself, and they are the rows that collapsed fastest.

It also explains something my own diary documents repeatedly: when I come off the protocol, these same symptoms return within weeks, in the same order. A signaling mechanism switches off when the signal stops. Rebuilt tissue would not behave that way. The reversibility is itself evidence about which mechanism is doing the work.

3. BHB unlocks genes, including the brain’s repair fertilizer (established in the laboratory; timeline in humans less clear)

Here BHB stops looking like a fuel altogether and starts looking like a hormone.

In 2013, a paper in Science showed that BHB is an endogenous inhibitor of class I histone deacetylases — HDACs, the enzymes that keep stretches of your DNA wound shut and therefore unreadable [3]. Inhibit them and specific genes become readable again. In that study, among the genes switched on were ones governing the cell’s defense against oxidative stress, and mice pre-treated with BHB were protected against a potent oxidative poison. Your own fasting chemistry, in other words, edits which of your genes are being read.

Then, in 2016, a paper in eLife connected that to the brain [4]. Exercise, it turns out, raises BHB in the hippocampus — the brain’s memory structure — and BHB there inhibits HDAC2 and HDAC3 at the promoter of the gene for BDNF, brain-derived neurotrophic factor: the protein that keeps neurons alive and grows new connections between them. Injecting BHB directly raised BDNF. Part of the reason exercise is good for your brain, on this evidence, is that exercise makes ketones.

A broad review in the Annual Review of Nutrition pulls this together: BHB is best understood as a signaling metabolite, linking the outside environment — what you ate, whether you fasted, whether you moved — to gene regulation, lipid metabolism, neuronal function, and metabolic rate [5]. I find this the most philosophically interesting fact in the whole field. The fuel is also the message.

4. Mitochondrial biogenesis — building new power plants (real, but almost certainly not what you are looking at here)

This is the mechanism people most want to invoke, and the one I want to be most careful about — including against my own commercial interest.

The evidence is real. In 2006, work published in Annals of Neurology found that a ketogenic diet increased mitochondrial biogenesis — the manufacture of new mitochondria — in the hippocampus of rats, and proposed this as part of why the diet controls seizures [6]. Later work described the machinery: upregulation of PGC-1α (the master switch for building new mitochondria), along with SIRT3 and UCP2, improving both the number and the performance of the cellular power plants [7].

So: does that explain my 622 to 268? Almost certainly not, and here is the reasoning.

Building new mitochondria is construction work — gene transcription, protein synthesis, assembly of organelles — and it operates on a timescale of months, not days. My scorecard covers forty-six days, much of it in the shallow end of the range while I was still climbing. Worse for the biogenesis story: my symptoms come back within weeks when I stop. Newly built mitochondria do not disassemble on that schedule. Signaling stops on that schedule.

My working judgment, stated so it can be checked against later columns: what you are looking at in this scorecard is roughly — fuel and de-inflammation, fast and reversible; a little epigenetic effect, in the middle; and structural rebuilding, barely started if at all. If real mitochondrial biogenesis is happening, it should show up as changes that appear later and persist longer — in the October, November, and December columns, and in what happens after Day 180. I have written the prediction down in advance, which is the only way a prediction counts.

Why the level and the duration both matter

One question follows naturally: if ketones do all this, why doesn’t everyone doing a standard low-carb diet report these results?

Because dose is not a detail. A typical ketogenic diet in practice produces BHB somewhere around 0.5 to 1.5 mmol/L — and in my clinic I have tested many patients doing everything they were told, eating protein and salad, no bread and no sugar, whose ketones were essentially zero. The therapeutic range I am describing is above 4.0, which is several multiples higher and requires far more precision to reach and hold. Whatever is dose-dependent about these mechanisms — and inflammasome inhibition and HDAC inhibition both are — a level that is three or four times lower is a genuinely different intervention.

Duration matters as much. My four-year diary shows something I only saw when I read it as data: single high readings protected me from almost nothing, while roughly a week of consecutive days above 4.0 changed my behavior entirely. I have written that finding up separately — it isn’t the reading, it’s the streak — and it is the practical companion to this article.

What didn’t move — and why that is the most important section

If every row had improved, you should distrust the entire document. Here is what did not.

My hearing: 6 in June, 6 now. Hair loss on my lower legs: 9 and 9. Brittle, discolored toenails: unchanged. New varicose veins: unchanged. Distance and near vision: unchanged — I am sixty, and presbyopia does not negotiate with ketones. Thigh muscle loss and the asymmetry between my legs: unchanged. Skin, hair and nails as a whole barely moved, 72 to 53, and most of that came from inflammatory items like scalp tenderness rather than from anything structural.

There is a pattern in the failures, and it is the same pattern the mechanisms predict. What moved was inflammatory and functional. What did not move was structural, degenerative, or cumulative — the results of decades of tissue change that no six weeks of anything reverses. Nails grow slowly. Hair follicles that have quit do not restart because the fuel improved. Sensorineural hearing loss is not an inflammatory condition. If my results had shown my hearing improving, the honest conclusion would have been that my scoring was unreliable.

One more honest complication: my sleep tracker’s percentages look worse in places — REM percentage down, interruptions up. But the device stops recording during long awakenings, which corrupts its own arithmetic, and the number I can verify — total time in deep sleep plus REM — is substantially higher than it was. When an instrument and lived experience disagree, examine what the instrument does badly before concluding anything about yourself. That is a rule I would apply to a patient’s wearable data, and I am applying it to my own.

How to read a study of one person — including this one

This is an n-of-1: a study with a single subject, no control group, no randomization, and an investigator who is also the participant and who sells a coaching program. Every one of those is a real limitation, and I would rather name them than have you find them.

The ways I could be wrong. Expectation is powerful, and I expected improvement — scoring my own symptoms is not blinded and cannot be. Regression to the mean is real: I began at a personal worst, and things at a worst tend to drift better regardless. Several confounders changed alongside ketosis: I lost weight, I began walking and eventually hiking, my sleep lengthened, and sunlight and season shifted from June to August. Any of those alone improves joints and mood. And a scoring scale in one person’s head is not a validated instrument, however carefully applied.

What still makes it worth something. The baseline was recorded before the outcome was known. The scoring was deliberately conservative. The predictions were written in advance and are being tested in public with dated entries. Some measures are instruments rather than impressions — glucose, heart rate variability. The failures are published alongside the successes. And most usefully: I have run this experiment on myself twenty-three previous times, in both directions, and the symptoms reliably return when the ketones fall. Repeated on-off-on within a single subject is the strongest design available to an individual — it is, in fact, what formal n-of-1 trials do deliberately.

What this is: one carefully documented case, consistent with well-established mechanisms, offered as a hypothesis worth testing — not as proof of anything general. What it is not: evidence that you will get these results, or medical advice of any kind.

Where this goes next

I re-score the full ninety-nine on the first of each month through December 13th, Day 180. The columns are already built and empty, and they will be published whatever they say. If the improvements plateau, you will see the plateau. If they reverse — if I break, as I have twenty-three times — you will see that too, on the same day it happens.

The specific thing I am watching for is the one I flagged above: changes that arrive late and hold. That would be the fingerprint of structural repair rather than signaling, and it is the difference between an intervention you must maintain forever and one that leaves you better than it found you. I do not know the answer. That is rather the point of doing it in public.

Bibliography

Each reference has a plain-English note about what it actually found, so you can judge the claims without a background in cell biology.

1. Youm YH, Nguyen KY, Grant RW, et al. “The ketone metabolite β-hydroxybutyrate blocks NLRP3 inflammasome-mediated inflammatory disease.” Nature Medicine. 2015;21(3):263–269.The foundational paper. BHB specifically blocks the NLRP3 inflammatory alarm — and the closely related ketone acetoacetate does not, nor do the related short-chain fats butyrate and acetate. It works by preventing potassium leaving the cell and stopping the alarm’s parts from clustering, and it does not run through the usual starvation pathways.
2. Goldberg EL, Asher JL, Molony RD, et al. “β-Hydroxybutyrate deactivates neutrophil NLRP3 inflammasome to relieve gout flares.” Cell Reports. 2017;18(9):2077–2087.Applies the mechanism to a real inflammatory disease. A ketogenic diet raising BHB reduced uric-acid-crystal gout inflammation by shutting down NLRP3-driven IL-1β release from neutrophils — without impairing the ability to fight bacterial infection. Targeted quieting, not blanket immune suppression.
3. Shimazu T, Hirschey MD, Newman J, et al. “Suppression of oxidative stress by β-hydroxybutyrate, an endogenous histone deacetylase inhibitor.” Science. 2013;339(6116):211–214.BHB is not only fuel: it inhibits class I HDACs, the enzymes that keep genes locked shut. Genes for handling oxidative stress switched on, and pre-treated mice survived a potent oxidative poison. This is the paper that made BHB a signaling molecule rather than a metabolic byproduct.
4. Sleiman SF, Henry J, Al-Haddad R, et al. “Exercise promotes the expression of brain derived neurotrophic factor (BDNF) through the action of the ketone body β-hydroxybutyrate.” eLife. 2016;5:e15092.Exercise raises BHB in the hippocampus, where it inhibits HDAC2 and HDAC3 at the BDNF gene — increasing the brain’s own growth-and-repair protein. Injecting BHB directly did the same thing. A mechanistic link between ketones and brain plasticity.
5. Newman JC, Verdin E. “β-Hydroxybutyrate: a signaling metabolite.” Annual Review of Nutrition. 2017;37:51–76.The comprehensive review. Catalogues BHB’s signaling roles at the cell surface and inside the cell — gene expression, lipid metabolism, neuronal function, metabolic rate — and argues these link environment to epigenetic regulation, with relevance to human disease and aging. The best single starting point for a serious reader.
6. Bough KJ, Wetherington J, Hassel B, et al. “Mitochondrial biogenesis in the anticonvulsant mechanism of the ketogenic diet.” Annals of Neurology. 2006;60(2):223–235.The origin of the mitochondrial-biogenesis claim: a ketogenic diet increased the building of new mitochondria in rat hippocampus, offered as part of why the diet stops seizures. Animal work, and — note — over a considerably longer timescale than six weeks.
7. Hasan-Olive MM, Lauritzen KH, Ali M, et al. “A ketogenic diet improves mitochondrial biogenesis and bioenergetics via the PGC1α-SIRT3-UCP2 axis.” Neurochemical Research. 2019;44(1):22–37.Describes the machinery behind the previous finding — PGC-1α is the master switch for building new mitochondria — and reports improvements in both the number and the performance of the cell’s power plants.
8. Abram SV, Kyner JM, Vu A, Naeem Z, Sethi S, Jacob MS, Fryer SL, Mathalon DH, Ford JM. “Metabolic improvements with a ketogenic diet correlate with symptom improvement in psychosis: a randomized controlled trial.” Schizophrenia Bulletin. 2026;52(4):sbag082.A randomized controlled trial in humans — 28 on a ketogenic diet versus 30 on their usual diet — showing improvements in metabolic health, cognition, and psychiatric symptoms. The finding that matters most here: improvement in depressive symptoms tracked with ketosis, not with weight loss, and weight loss was unrelated to the metabolic and symptom improvements. Evidence that the ketones, not the pounds, are doing the work.

Educational only, and not medical advice. This is a documented self-experiment in one person. Therapeutic ketosis can interact with prescription medication — particularly for diabetes and heart failure — and is not appropriate for everyone, including people with kidney disease, liver disease, cardiac arrhythmia, or a history of eating disorders. Talk to your own physician before making any changes.