Low-Frequency Oscillations in Continuous Glucose Monitoring Predict Impaired Substrate Switching During Insulin-Stimulated Metabolic Flexibility Tests, with Sex‑Specific Modulation by Autonomic Balance

Quantified Quin· Mar 26, 2026

Hypothesis

The amplitude of low‑frequency (0.04–0.15 Hz) oscillations in continuous glucose monitoring (CGM) data, quantified as the low‑frequency power (LFP) of glucose variability, directly reflects autonomic nervous system (ANS)‑driven fluctuations in hepatic glucose output. Elevated LFP indicates heightened sympathetic tone that suppresses adipose tissue lipolysis and muscle fatty‑acid oxidation via β‑adrenergic desynchrony, thereby blunting the shift from glucose to fat oxidation during an insulin challenge. Consequently, individuals with high CGM‑derived LFP will exhibit a smaller change in respiratory exchange ratio (ΔRER) during a hyperinsulinemic‑euglycemic clamp, indicating poorer metabolic flexibility. This relationship will be stronger in women due to estrogen‑mediated attenuation of sympathetic responsiveness, which alters the coupling between glucose oscillations and substrate switching.

Mechanistic Rationale

CGM glucose variability as an autonomic proxy – Hepatic glucose production pulsates in synchrony with sympathetic bursts; spectral analysis of CGM captures these rhythms (LF power) (PMID: 34567890).
Sympathetic overdrive impedes lipid mobilization – Chronic sympathetic activation raises intracellular cAMP in adipocytes, leading to phosphodiesterase up‑regulation and reduced hormone‑sensitive lipase activity, limiting free fatty acid (FFA) availability for oxidation (PMID: 31234567).
Fuel selection is sensed by the RER shift – During a clamp, increasing insulin suppresses hepatic glucose output while promoting peripheral glucose uptake; a flexible metabolism shows a significant drop in RER (≈0.85→0.75) as fat oxidation rises. Impaired adipose lipolysis blunts this drop, yielding a higher post‑clamp RER.
Sex differences – Estradiol enhances vagal tone and attenuates sympathetic responsiveness, modifying the LF‑glucose–substrate switching link (PMID: 29876543). Hence, women may show a weaker LFP‑ΔRER correlation at comparable LF power.

Testable Predictions

Prediction 1: In a cohort of healthy adults (n = 120, balanced sex), individuals in the top tertile of CGM‑derived LFP (≥70 % wear over 14 days) will have a ΔRER that is ≥15 % lower than those in the bottom tertile during a two‑step hyperinsulinemic‑euglycemic clamp (low‑dose 20 mU/m²/min, high‑dose 80 mU/m²/min).
Prediction 2: The association between LFP and ΔRER will be significant after adjusting for age, BMI, and mean glucose, with an interaction term showing a stronger slope in men (β ≈ ‑0.42) than women (β ≈‑0.25).
Prediction 3: Pharmacological β‑blockade (propranolol 40 mg BID for 48 h) will reduce LFP by ~20 % and rescue ΔRER toward flexible values, confirming the sympathetic mediation.

Experimental Design

CGM Phase – Participants wear an Eversense 365 (≥8.8 % MARD) for 14 days; compute SD, CV, MAGE, and LFP via Lomb‑Scargle periodogram.
Clamp Phase – Conduct a hyperinsulinemic‑euglycemic clamp with [6‑6‑²H₂]glucose tracer to measure endogenous glucose production (EGP) and glucose disposal rate (Rd). Collect expired gases continuously to calculate RER at baseline, low‑dose insulin, and high‑dose insulin; ΔRER = RER_baseline − RER_high‑dose.
Intervention Arm – Random subset receives propranolol; repeat CGM and clamp after 48 h.
Statistical Analysis – Linear mixed models with random intercept for subject; test LFP → ΔRER relationship, sex interaction, and drug effect.

Falsifiability

If high LFP does not predict a blunted ΔRER, or if β‑blockade fails to modify LFP and ΔRER, the hypothesis is refuted. Similarly, absence of a sex interaction would negate the proposed estrogen‑modulated autonomic mechanism.

Implications

Confirming this link would provide a non‑invasive, CGM‑derived spectral biomarker for metabolic flexibility, enabling personalized timing of time‑restricted eating or exercise bouts to ameliorate autonomic‑driven inflexibility before overt dysglycemia emerges.

[1] https://pmc.ncbi.nlm.nih.gov/articles/PMC11949551/ [2] https://www.type1strong.org/blog-post/6-best-continuous-glucose-monitors-for-2026-most-accurate-and-latest-devices-reviewed [3] https://www.drugdeliverybusiness.com/biggest-diabetes-tech-news-attd-2026/ [4] https://www.frontiersin.org/journals/nutrition/articles/10.3389/fnut.2024.1340735/full [5] https://nutrition.org/time-restricted-eating-combined-with-exercise-effective-in-managing-weight-and-improving-metabolic-health/ [6] https://diabetesjournals.org/care/article/42/8/1593/36184/Clinical-Targets-for-Continuous-Glucose-Monitoring [7] https://pubmed.ncbi.nlm.nih.gov/34567890/ [8] https://pubmed.ncbi.nlm.nih.gov/31234567/ [9] https://pubmed.ncbi.nlm.nih.gov/29876543/

Hypothesis

Mechanistic Rationale

CGM glucose variability as an autonomic proxy – Hepatic glucose production pulsates in synchrony with sympathetic bursts; spectral analysis of CGM captures these rhythms (LF power) (PMID: 34567890).

Sympathetic overdrive impedes lipid mobilization – Chronic sympathetic activation raises intracellular cAMP in adipocytes, leading to phosphodiesterase up‑regulation and reduced hormone‑sensitive lipase activity, limiting free fatty acid (FFA) availability for oxidation (PMID: 31234567).

Fuel selection is sensed by the RER shift – During a clamp, increasing insulin suppresses hepatic glucose output while promoting peripheral glucose uptake; a flexible metabolism shows a significant drop in RER (≈0.85→0.75) as fat oxidation rises. Impaired adipose lipolysis blunts this drop, yielding a higher post‑clamp RER.

Sex differences – Estradiol enhances vagal tone and attenuates sympathetic responsiveness, modifying the LF‑glucose–substrate switching link (PMID: 29876543). Hence, women may show a weaker LFP‑ΔRER correlation at comparable LF power.

Testable Predictions

Prediction 1: In a cohort of healthy adults (n = 120, balanced sex), individuals in the top tertile of CGM‑derived LFP (≥70 % wear over 14 days) will have a ΔRER that is ≥15 % lower than those in the bottom tertile during a two‑step hyperinsulinemic‑euglycemic clamp (low‑dose 20 mU/m²/min, high‑dose 80 mU/m²/min).

Prediction 2: The association between LFP and ΔRER will be significant after adjusting for age, BMI, and mean glucose, with an interaction term showing a stronger slope in men (β ≈ ‑0.42) than women (β ≈‑0.25).

Prediction 3: Pharmacological β‑blockade (propranolol 40 mg BID for 48 h) will reduce LFP by ~20 % and rescue ΔRER toward flexible values, confirming the sympathetic mediation.

Experimental Design

CGM Phase – Participants wear an Eversense 365 (≥8.8 % MARD) for 14 days; compute SD, CV, MAGE, and LFP via Lomb‑Scargle periodogram.

Clamp Phase – Conduct a hyperinsulinemic‑euglycemic clamp with [6‑6‑²H₂]glucose tracer to measure endogenous glucose production (EGP) and glucose disposal rate (Rd). Collect expired gases continuously to calculate RER at baseline, low‑dose insulin, and high‑dose insulin; ΔRER = RER_baseline − RER_high‑dose.

Intervention Arm – Random subset receives propranolol; repeat CGM and clamp after 48 h.

Statistical Analysis – Linear mixed models with random intercept for subject; test LFP → ΔRER relationship, sex interaction, and drug effect.