Keto-Mojo GK+: Capillary Verification & GKI

In the Guesswork Era, metabolic monitoring ended at fasting glucose — a single capillary check that measured the body’s baseline state while ignoring everything that happened between meals, overnight, and under fasting stress. At Biohack Forge , we recognise that a continuous glucose monitor provides trend velocity, but interstitial fluid is not blood — and blood is not the complete metabolic picture. The Keto-Mojo GK+ was deployed to provide two things the Dexcom ONE+ cannot: capillary ground truth at moments of rapid metabolic shift, and Beta-Hydroxybutyrate quantification during therapeutic fasting. Together, they constitute a dual-layer intelligence system that neither device can replicate alone.

01. The Objective

The Dexcom ONE+ (Briefing 01.1) is the trend engine — a continuous stream of interstitial glucose data that reveals amplitude, frequency, and circadian patterning across a 24-hour metabolic timeline. But interstitial fluid carries a 5–25 minute physiological lag behind capillary blood glucose. During periods of rapid metabolic transition — a fast-rising postprandial spike, a post-exercise glucose drop, or the early hours of an extended fast — this lag is not a rounding error. It is a structural data gap that can cause algorithmic smoothing to obscure the true peak or nadir.

The Keto-Mojo GK+ was deployed to fill two gaps the CGM cannot close. First: to provide capillary blood glucose as a ground-truth reference layer during the prediabetes recovery protocol, ensuring that CGM-derived protocol decisions were anchored to real blood values. Second: to quantify Beta-Hydroxybutyrate (BHB) during a 72-hour water fast — enabling Glucose Ketone Index (GKI) tracking as a proxy for the metabolic state associated with mTOR suppression and cellular repair signalling.

The Keto-Mojo GK+ was deployed to answer one question: does the CGM’s interstitial data hold up against capillary ground truth at the moments that matter most — and can BHB quantification map the metabolic states that glucose monitoring cannot reach?

02. The Catalyst

Two distinct triggers initiated this audit, separated by approximately 60 days within the same recovery arc.

Trigger 1 — CGM Accuracy Validation: Following deployment of the Dexcom ONE+ to investigate a prediabetic HbA1c of 42 mmol/mol (6.0%), it became operationally necessary to validate the CGM’s postprandial readings against capillary ground truth. The Dexcom ONE+ Briefing documented CGM accuracy at $\pm0.3\ mmol/L$ average deviation — but this cross-validation data came from the Keto-Mojo GK+.

Trigger 2 — Extended Fasting Protocol (Day 60): With postprandial architecture stabilised and HbA1c trajectory confirmed as reversed (progressing toward the 37 mmol/mol / 5.5% reading documented), a 72-hour water fast was initiated to audit Pillar 04: Systemic Defense & Repair mechanisms — specifically, the metabolic state associated with autophagy signalling and insulin resensitisation. The Keto-Mojo GK+ was the only instrument capable of tracking the Glucose Ketone Index required for this phase.

A note on HbA1c values across this briefing series: the 37 mmol/mol (5.5%) outcome documented in Decom One+ represents the 90-day CGM window endpoint.

03. Hardware Architecture

Upgrade Delta

The Keto-Mojo GK+ occupies a distinct device category from both consumer ketone strips and clinical laboratory analysers. Consumer ketone strips (e.g., Ketostix urinary strips) measure acetoacetate — a ketone body that does not correlate reliably with BHB at varying hydration states and provides no quantitative precision. Finger-prick BHB strips from competing brands (Precision Xtra, Abbott) measure the same biomarker but lack Bluetooth data integration, making longitudinal pattern analysis manual and error-prone.

The GK+ combines capillary blood glucose and BHB on a single platform, with $0.1\ mmol/L$ BHB precision — the resolution required for meaningful GKI calculation. At $\pm0.1\ mmol/L$ , the difference between GKI 0.8 and GKI 1.1 is resolvable; on a strip with $0.5\ mmol/L$ precision, it is not. For a protocol where the GKI threshold carries interpretive weight, precision is not a preference — it is a prerequisite.

Bluetooth integration enables automatic data logging to the Keto-Mojo app, producing a timestamped record that cross-references against Dexcom Clarity exports. This is the operational basis for the dual-layer validation methodology in Section 04.

Form Factor and Protocol Compatibility

The spot-check form factor is the correct tool for this use case. Unlike the Dexcom ONE+, which requires 24/7 adhesion across Pillar 02 training loads, the GK+ is deployed at defined metabolic junctions — pre-meal, postprandial peak, post-exercise, fasted — without interfering with any other protocol element. The lancing device is compatible with forearm and palm sites, allowing fingertip rotation across extended testing periods without site fatigue.

During the 72-hour fast, the GK+ was the only active monitoring instrument.

Cross-Validation Method

The GK+ served as the reference standard for Dexcom ONE+ validation throughout the prediabetes recovery phase. Simultaneous testing was conducted at:

Fasted baseline (06:00–06:30)
60 minutes post-meal (postprandial peak window)
Immediately post-exercise (0–15 minutes after Pillar 02 sessions)

Average deviation between GK+ capillary and Dexcom ONE+ interstitial readings: $\pm0.3\ mmol/L$ . This figure originates from this cross-validation dataset. The GK+ itself was not validated against a laboratory venous blood draw in this audit — a limitation acknowledged in Section 04.

04. Protocol Methodology

Duration: 90-day prediabetes recovery phase (BGM/CGM cross-validation) + 72-hour water fast (GKI audit).

Data Capture Windows:

BGM/CGM Cross-Validation Phase:

Fasted baseline: 06:00–06:30 daily (simultaneous GK+ and Dexcom reading)
Postprandial peak: 60 minutes post each primary meal (timed to align with typical CGM peak window)
Post-exercise: 0–15 minutes after all Pillar 02 training sessions

72-Hour Extended Fast:

BHB and glucose: every 12 hours from fast initiation (19:00 Day 1)
Additional readings at subjective symptoms (dizziness, significant hunger, cognitive shift)
GKI calculated from simultaneous glucose and BHB readings at each checkpoint

Cross-Pillar Integration:

Pillar 01 (Metabolic Intelligence): GK+ data served as the calibration anchor for all CGM-derived protocol decisions across the 90-day window.
Pillar 04 (Systemic Defense & Repair): GKI trajectory during the 72-hour fast was used to assess the metabolic state associated with mTOR suppression and autophagy signalling.

Interventions Tested:

BGM/CGM simultaneous cross-validation to identify interstitial lag magnitude and algorithmic smoothing artefacts
BHB tracking of nominally keto-compatible processed foods to detect insulin-mediated ketone suppression without corresponding glucose excursion
72-hour water fast with 12-hourly GKI tracking to map the metabolic switch timeline

Confounders Acknowledged:

The GK+ was not validated against a laboratory venous reference within this audit. The device carries CE/FDA clearance and published accuracy data, but within-audit laboratory confirmation was not performed. The BGM readings are treated as protocol-grade, not clinical-grade.
The 72-hour fast was conducted during a period of low training load (Pillar 02), which may have accelerated glycogen depletion relative to a higher-activity baseline.
Hydration status was monitored subjectively rather than by plasma osmolality — a confounder for both BHB concentration readings and GKI calculation during prolonged fasting.

05. Metadata Findings

A. The Interstitial Lag: Where the CGM Is and Is Not Reliable

Simultaneous BGM/CGM readings across the 90-day audit confirmed the theoretical 5–25 minute interstitial lag in practice. At the postprandial peak window — the point of maximum clinical relevance for the prediabetes protocol — the Dexcom ONE+ consistently reported values $0.3–0.8\ mmol/L$ below the concurrent GK+ capillary reading, with the gap widening during rapid glucose rise phases.

The practical consequence: a CGM reading of $8.2\ mmol/L$ at 45 minutes post-meal corresponded to a simultaneous capillary reading of $8.7–9.0\ mmol/L$ . For a protocol calibrating the threshold between an acceptable postprandial excursion ( $<8.0\ mmol/L$ capillary) and a spike requiring dietary intervention, this difference is operationally significant. The CGM would have classified the meal as borderline-acceptable; the BGM classified it as a protocol breach.

The mechanism: the 5–25 minute lag is a physiological property of interstitial fluid diffusion kinetics — CGM sensors sample the extracellular matrix, not plasma, and during a rapid glucose rise, capillary concentration consistently leads interstitial concentration until the diffusion gradient equalises.

Protocol decision: CGM postprandial readings above $7.5\ mmol/L$ were cross-referenced against a simultaneous GK+ capillary check before any dietary modification was logged as “passed.” This raised the effective intervention threshold by approximately $0.4–0.6\ mmol/L$ relative to CGM-only monitoring and tightened the carbohydrate elimination criteria in Interventions 3 and 4 of the Dexcom protocol.

B. BHB as an Insulin Signal: The Hidden Suppression Phenomenon

Glucose monitoring — whether CGM or BGM — cannot detect an insulin response that does not produce a significant glucose excursion. For nominally keto-compatible processed foods (certain protein bars, commercially marketed “keto” bread products, sugar alcohol-sweetened confectionery), this creates a diagnostic blind spot: the product does not spike glucose, the CGM reports nothing unusual, and the protocol proceeds on the assumption that the metabolic impact is negligible.

BHB tracking identified a different signal. Several commercially available keto-marketed products produced a measurable ketone suppression — BHB declining from $0.8–1.1\ mmol/L$ (light ketosis) to $0.2–0.3\ mmol/L$ within 90 minutes of consumption, with no corresponding glucose excursion above $7.0\ mmol/L$ . The mechanism: insulinogenic amino acid profiles and the insulinotropic effects of certain sugar alcohols can elicit an insulin response sufficient to suppress hepatic ketogenesis without producing a glucose spike detectable by BGM or CGM.

Forge Note: This finding has direct implications for Pillar 01 dietary architecture. The absence of a glucose spike is not evidence of metabolic neutrality. A product can be glycemically inert and still disrupt fat-oxidation architecture via insulin-mediated ketone suppression. Without concurrent BHB tracking, this signal is invisible.

C. The 72-Hour Fast: GKI Trajectory and Metabolic Switch

Critical framing note: The GKI was developed by Thomas Seyfried (Boston College) primarily in the context of cancer metabolic therapy, not as a validated autophagy biomarker for healthy individuals. The mechanistic chain — low glucose → low insulin → mTOR suppression → autophagy upregulation — is coherent and supported by animal model and in vitro data. However, no randomised controlled trial has validated specific GKI thresholds against direct autophagy markers (LC3-II, p62, beclin-1) in healthy humans. GKI is used here as a mechanistic proxy only.

The 72-hour water fast was initiated from a metabolic baseline of $5.1\ mmol/L$ fasting glucose and $0.8\ mmol/L$ BHB — a GKI of 6.4, consistent with a glucose-dependent metabolic state. The GKI progression across the fast:

$GKI = \frac{Glucose\ (mmol/L)}{Ketones\ (mmol/L)}$

Time Point	Glucose	BHB	GKI	Metabolic State
0h (Baseline)	$5.1\ mmol/L$	$0.8\ mmol/L$	6.4	Glucose-dependent
24h	$4.0\ mmol/L$	$0.8\ mmol/L$	5.0	Metabolic transition
36h	$5.8\ mmol/L$	$2.8\ mmol/L$	2.1	Early therapeutic ketosis
48h	$3.9\ mmol/L$	$3.0\ mmol/L$	1.3	Therapeutic ketosis
60h	$3.8\ mmol/L$	$4.5\ mmol/L$	0.8	Deep ketosis
72h	$3.5\ mmol/L$	$3.9\ mmol/L$	0.9	Deep ketosis

The metabolic switch — defined here as the point where fat-derived ketone production exceeded the rate of glucose-dependent energy demand — was confirmed at approximately hour 36. This is consistent with glycogen depletion timelines for an individual with a lower carbohydrate baseline ( $<100g$ daily) and moderate Pillar 02 training history.

At 60 hours, GKI 0.8 places the metabolic state in the range associated with significant mTOR suppression in the mechanistic literature. Subjectively: cognitive clarity was high, hunger had subsided completely by hour 20, and no hypoglycaemic symptoms were recorded — consistent with the metabolic flexibility established during the 90-day Dexcom protocol.

Protocol decision: The 72-hour fast confirmed metabolic flexibility as operationally validated. The extended fast protocol is retained as a quarterly Pillar 04 intervention, with GKI monitoring at 12-hourly intervals as the decision instrument for safe continuation and termination.

D. Fasting Safety Architecture

The GK+ served as the objective safety instrument during the 72-hour fast — the function the raw notes appropriately describe as a “safety dashboard.” The critical safety parameter in extended fasting is not hunger or fatigue (both subjective and unreliable) but the inverse relationship between declining glucose and rising ketones.

A falling glucose reading in isolation can represent hypoglycaemia — a genuine safety concern. The same falling glucose reading paired with rising BHB confirms the metabolic switch: the brain and peripheral tissues are being supplied by ketones as glucose declines to its fasted homeostatic range ( $3.5–4.0\ mmol/L$ ). Without concurrent BHB data, a glucose reading of $3.5\ mmol/L$ at hour 72 could not be distinguished from early hypoglycaemia. With BHB at $3.9\ mmol/L$ , it is unambiguously the expected substrate shift of deep ketosis.

The mechanism: when hepatic glycogen is depleted and insulin is suppressed, the liver increases fatty acid oxidation and ketogenesis via AMPK activation — BHB enters systemic circulation as the primary fuel substrate, and blood glucose stabilises at a lower homeostatic setpoint rather than continuing to fall, which is the pathological trajectory of true hypoglycaemia.

Protocol decision: No extended fast ( $\geq24$ hours) is conducted without GK+ BHB monitoring. Glucose-only fasting safety assessment is classified as insufficient for the Forge Protocol.

06. Consensus 14 Outcomes

Marker	Baseline	Optimised	Status
Fasting Glucose	$5.2\ mmol/L$	$4.8\ mmol/L$	▲ Optimised
HbA1c	42 mmol/mol	37 mmol/mol	▲ Reversed
Beta-Hydroxybutyrate (fasted)	$0.8\ mmol/L$	$3.9\ mmol/L$ (72h fast)	▲ Validated
GKI (72h fast endpoint)	6.4	0.9	▲ Therapeutic threshold reached
Metabolic Flexibility	Unconfirmed	Confirmed (GKI < 1.0 sustained)	▲ Validated

Forge Verdict: The Keto-Mojo GK+ resolved two data gaps that the CGM cannot close: the interstitial lag that was masking true postprandial peaks during prediabetes recovery, and the BHB suppression signal that identified insulin-mediated metabolic disruption in foods that glucose monitoring alone would have approved. The 72-hour fast confirmed that the metabolic flexibility built during the 90-day Dexcom protocol is operationally real — the hardware switched fuel sources without hypoglycaemic crisis, sustained GKI < 1.0 for the terminal 24 hours, and recovered to baseline glucose within 4 hours of refeeding. The dual-device stack is not redundancy. It is depth.

Protocol Status: ADOPTED

The Keto-Mojo GK+ carries Adopted status within the Biohack Forge Protocol 1.0. It is designated as the mandatory reference instrument for any individual running concurrent CGM monitoring, and as the primary safety and efficacy tracking tool for all extended fasting protocols ( $\geq24$ hours) within Pillar 04.

It does not replace continuous glucose monitoring — it anchors it. The Dexcom ONE+ provides the trend; the GK+ provides the ground truth when the trend is making a protocol-critical call.