One participant. Resting heart rate 65 to 46 beats per minute.
Weight from over 90 kilograms to 72. Metabolic switching from more than 24 hours down to 16.5, against a 36 to 72 hour range typical for his age.
The participant is Dean Ho, 47, who directs the Institute for Digital Medicine at NUS Medicine and who designed the study he is the subject of. It is called DELTA. It went into PLOS One on 12 August and it has not stopped running.
The protocol is specific, which is the part worth copying and the part nobody quotes.
Feeding window of four hours, 11am to 3pm, with a minimum 20-hour fast. Multiple 48-hour water-only fasts.
Strength training daily, one muscle group at a time, plus two or three Norwegian-protocol HIIT sessions a week. Mediterranean-style food, and nothing to drink beyond water, electrolytes, black coffee and unsweetened black tea.
Bedtime advanced two and a half hours, from midnight to 9.15pm.
Sleep went from 372.3 minutes a night at retrospective baseline to 454.6 at study start and 493.3 at the end. That is a little over two extra hours in bed, held for eight months.
The number that will travel is the one the paper is least able to defend.
The number that will travel is the one the paper is least able to defend.
Estimated biological age comes in at 31.8 against a chronological 45, produced by an ordinary least-squares model over ten health metrics fitted on a general population sample. The authors write that the estimate has no external validation cohorts and no pre-specified performance criteria, and that it should be read inside the hypothesis-generating scope of an N-of-1 design.
They go further. With no untreated control, apart from the sleep assessment, they make no efficacy claims for the protocol at all.
Grade it Level 3. One subject, no control arm, the protocol’s designer as the participant, and a retrospective baseline gathered before anyone was holding season or workload steady. The paper says all of this in its own limitations section, which is more than most of the coverage will.
What survives the discount is the cheaper half of the study.
Sleep is the one endpoint DELTA did test against its own baseline, and it moved 121 minutes. Resting heart rate fell 19 beats. Neither figure needs a model to compute, and neither needs a panel to confirm.
The cost is denominated in hours. A four-hour window in which you are allowed to eat, ninety minutes of training before the day starts, and an evening that ends at a quarter past nine.
You are buying the biomarker with the part of the day most people use to have a life.
You are buying the biomarker with the part of the day most people use to have a life.
That is the disclosure this study makes better than any clinic panel does, because the intervention is legible line by line. The trade is visible, the instruments are a wrist watch and a bedtime, and the strongest results came off the two levers that cost nothing to pull and everything to keep pulling.1
Of every number DELTA published, the two that hold up are the two a man could read off his own wrist: how long he sleeps, and how slowly his heart beats when he is doing nothing at all.
Footnotes
-
The biological-age figure came out of a custom AI copilot built for the study. The Garmin on his wrist came out of a shop. ↩