You already track glucose, HRV, and sleep. The gap isn't data—it's a framework for deciding which longevity claims are mechanism and which are marketing.
This article is educational and not medical advice. Nothing here recommends a specific drug, dose, or protocol; those decisions belong to an independent licensed provider who has reviewed your labs and history.
What autophagy actually is
Autophagy—literally "self-eating"—is a conserved cellular recycling process. When a cell is stressed or nutrient-deprived, it packages damaged proteins and organelles into double-membrane vesicles (autophagosomes) that fuse with lysosomes for degradation and reuse. The 2016 Nobel Prize in Physiology or Medicine went to Yoshinori Ohsumi for mapping the core machinery in yeast, which established autophagy as a fundamental quality-control system [1].
The key regulatory node is mTOR (mechanistic target of rapamycin). When nutrients—especially amino acids and insulin signaling—are abundant, mTOR is active and suppresses autophagy. When energy is low, the sensor AMPK activates and mTOR quiets down, which lifts the brake on autophagy [2]. This is the mechanistic reason fasting is linked to autophagy at all: it shifts the mTOR/AMPK balance.
Here's the honest caveat an engineer should hold onto: almost all of the clean, causal autophagy data comes from yeast, worms, mice, and cell culture. Measuring autophagy directly in a living human is genuinely hard. There's no wearable for it, no validated blood marker you can export to a spreadsheet, and the timing curves people quote ("autophagy peaks at hour X") are largely extrapolated from animal models, not measured in humans [1][2].
Source: [1] The Nobel Prize in Physiology or Medicine 2016 (Yoshinori Ohsumi, autophagy), [2] AMPK and mTOR in Cellular Energy Homeostasis and Autophagy (Cell Metabolism / NIH PMC)
Where fasting evidence is solid—and where it thins out
Fasting and time-restricted eating do produce measurable metabolic effects in humans. A widely cited review in *The New England Journal of Medicine* summarizes intermittent fasting's effects on the "metabolic switch" to ketone-based fuel and downstream signaling [3]. That's real physiology you can partly see on a CGM: narrower glucose excursions, lower fasting glucose in some people.
What thins out is the leap from "fasting changes metabolic signaling" to "fasting produces a specific, dose-like amount of autophagy that extends my healthspan." That bridge isn't built in humans. Randomized human trials of time-restricted eating have shown mixed results on body composition and cardiometabolic markers, and a widely discussed randomized trial found time-restricted eating without calorie counting produced results similar to unrestricted meal timing when total calories were matched [4]. The signaling is real; the magnitude and clinical payoff in humans remain uncertain.
For a self-quantifier, the takeaway is precision, not dismissal: fasting is a legitimate lever on metabolic signaling. It is not a validated dial for a measurable quantity of autophagy.
The peptide question, read skeptically
This is where forum enthusiasm most outruns published human evidence. "Peptide" is a category, not a claim—it just means a short chain of amino acids. Some peptides are well-studied FDA-approved drugs; many circulating in the biohacking world are not, and the gray-market supply chain is a genuine purity and identity problem, not a paranoid one.
A few grounding facts:
- Regulatory reality. In 2023, the FDA reviewed several substances proposed for compounding and placed a number of peptide-related ingredients into a category flagged for significant safety concerns or insufficient data, restricting their use in compounding [5]. That is a documented signal that "available from a compounding source" does not mean "characterized and safe."
- Purity is not guaranteed by a website. Research-use-only vials sold online are not held to the identity, sterility, and potency standards of dispensed medications, and independent analyses have repeatedly found mislabeled or contaminated products in unregulated supplement and peptide markets [5].
- Human outcome data is often thin. Many peptides with enthusiastic forum followings have mechanistic or animal rationale but little or no rigorous human trial evidence for the longevity outcomes people assume. Mechanism is a hypothesis, not a result.
Compounded medications are not reviewed or approved by the FDA for safety, effectiveness, or quality. Compounded products are not equivalent to or interchangeable with any FDA-approved brand-name drug. Availability varies by state. Whether any prescription is appropriate is a decision only an independent licensed provider can make after evaluating you—it is never guaranteed.
Source: [5] FDA: Compounding and the FD&C Act (Bulk Drug Substances / 503A categories)
How a physician reads your self-quantified data
Your CGM traces, HRV trends, and sleep exports are useful context, but a provider anchors them to validated labs before drawing conclusions. A few of the markers that carry more interpretive weight than any wearable:
- HbA1c and fasting glucose/insulin for the metabolic picture your CGM only hints at. The American Diabetes Association publishes the reference cut points clinicians actually use [6].
- A full lipid panel, read in the context of cardiovascular risk frameworks from the ACC/AHA [7].
- Baseline organ-function and hormone panels relevant to whatever you're considering, so there's a real "before" to compare against.
The clinical value isn't the raw number—it's the trajectory against a validated baseline, plus context (medications, symptoms, family history) that no sensor captures. A good provider treats your spreadsheet as a hypothesis generator and the labs as the arbiter. That's also how side effects and off-target changes get caught early, which matters more with anything experimental.
% HbA1c · marker = Diabetes threshold
Source: [6] American Diabetes Association: Standards of Care — Classification and Diagnosis
A cleaner mental model
1. Mechanism (plausible in cells/animals) → signaling change (sometimes measurable in humans) → clinical outcome (often unproven). Know which tier a claim lives in before you act on it.
2. Measured beats extrapolated. Trust the CGM trend and the lab draw over an animal-derived timing curve.
3. Supply chain is part of the molecule. An uncharacterized vial isn't the same intervention as a dispensed, quality-controlled medication—even if the label says the same word.
Where Velri fits
Velri is a technology and coordination company—not a medical practice. Velri doesn't provide care or dispense medication. What it can coordinate is the structure this article argues for: baseline lab work, a visit with an independent, licensed provider from the affiliated provider groups who will actually engage with your goals and data, and—only if that provider determines it's appropriate and writes a prescription—fulfillment through an independent, licensed pharmacy. The clinical judgment, and any decision to prescribe, stays entirely with the provider. This is education to help you ask sharper questions, not a promise of any specific treatment.
