What a Fasting Glucose of 106 Tells You

Oak
glucose cgm prediabetes biomarkers hba1c
Long-legged translucent creatures graze across a faint glowing green line on bioluminescent tidal flats at dusk.

A fasting glucose of 106 mg/dL falls in the 100–125 mg/dL band that the American Diabetes Association calls impaired fasting glucose, or prediabetes. That is a statement about a population-level risk threshold, not a description of your physiology. The practical reading is that 106 sits close enough to the cut point that the measurement itself is the first thing to question: day-to-day biological variation in fasting glucose is roughly 5–8 mg/dL within the same person, and a tube of blood that sits at room temperature before the plasma is separated will drift upward because red cells keep consuming glucose at a few percent per hour. A single 106 is a reason to measure again properly and to add two or three orthogonal measurements, not a reason to conclude anything yet. If a repeat value and an HbA1c both land in the impaired range, that is a conversation with a physician, and this page does not substitute for it.

Where the thresholds come from and how tight they are

The diagnostic cut points are administrative decisions about where risk curves bend, and they have moved. The fasting threshold for impaired fasting glucose was 110 mg/dL until 2003, when the ADA lowered it to 100. Nothing about human metabolism changed. The 100 mg/dL line means a large fraction of adults sit near it, which is why the boundary between “97 fasting glucose” and “106 fasting glucose” carries far less information than the numbers suggest.

Three tests define the same territory from different angles. Fasting plasma glucose reflects hepatic glucose output overnight and the insulin required to suppress it. The two-hour value in a 75 g oral glucose tolerance test reflects peripheral disposal capacity and first-phase insulin secretion. HbA1c, glycated hemoglobin, integrates exposure over the lifespan of your red cells, weighted toward the last four to six weeks. They disagree in individuals more often than they agree. Someone with a fasting glucose of 121 and a normal OGTT is a different phenotype from someone with a fasting glucose of 96 and a two-hour value of 165, and the second person carries more cardiovascular risk despite the reassuring fasting number.

The relationship people ask about, “what is my A1C if my blood sugar is 109,” is a population regression, not a conversion. Estimated average glucose maps to HbA1c by the ADAG equation (average glucose in mg/dL ≈ 28.7 × A1c − 46.7), so an A1c of 5.7% corresponds to a mean glucose near 117. A single fasting reading is not a mean glucose, and hemoglobin glycation rates vary between people for reasons unrelated to glucose, including red cell lifespan, anemia, and hemoglobin variants. Use A1c as its own measurement, not as a translation of a fingerstick.

Prediabetes is also not a one-way door. Regression from impaired fasting glucose or impaired glucose tolerance back to normal glucose regulation is well documented, occurs in a substantial minority of people over one to three years of follow-up in the large prevention trials, and predicts lower subsequent diabetes incidence than staying in the intermediate category.1 What drives regression is mostly weight and fitness, which is a matter for your clinician, not for this page.

The measurements we would order around a 106

If we saw a 106 in our own data, we would assemble a five-part picture rather than repeat the same test.

Repeat the fasting plasma glucose under controlled conditions. Twelve hours fasted, water only, no exercise the prior evening, blood drawn in a gray-top sodium fluoride/potassium oxalate tube or spun within thirty minutes. Two values a week apart tell you your own within-person variance, which you need before reading any single point.

Get an HbA1c and, if available, fasting insulin measured in the same draw. Fasting insulin lets you compute HOMA-IR ((fasting insulin in µU/mL × fasting glucose in mg/dL) / 405) and QUICKI, crude but useful indices of hepatic insulin sensitivity. A fasting glucose of 106 with an insulin of 4 µU/mL means something different from 106 with an insulin of 18.

Run a 75 g OGTT with draws at 0, 30, 60, 90, and 120 minutes if you can get them. The shape matters. A monophasic curve with a late peak is associated with worse beta-cell function than a biphasic curve with the same two-hour value.

Wear a continuous glucose monitor for fourteen days. CGM measures interstitial fluid glucose, not plasma glucose, and reports a value every five minutes. Professional guidance has moved toward using CGM metrics in their own right rather than as a surrogate for laboratory glucose.23 The two most informative summaries for someone in your position are time in a tight range (percentage of readings between 70 and 120 mg/dL) and glucose variability expressed as the coefficient of variation, standard deviation divided by mean.4

Finally, look at the genome. Fasting glucose is polygenic with modest per-variant effects, so a polygenic score is weakly informative, but monogenic causes matter here. Heterozygous GCK loss-of-function variants (MODY2, glucokinase-related maturity-onset diabetes of the young) raise the glucose set point and produce lifelong fasting values in the 100–145 mg/dL range with an A1c usually below 7.5% and little progression. A 106 that never moves, with a family history of the same, is worth screening for. Check GCK, HNF1A, and HNF4A in your VCF and run the coding variants through a clinical-grade annotation. A likely-pathogenic hit changes how a clinician interprets every glucose number you will ever produce, which is exactly why it is a clinician’s call and not yours.

Reading your own CGM export

Consumer and prescription sensors export CSV files that you can analyze directly. Dexcom’s Clarity export gives columns for timestamp, event type, and glucose value in mg/dL with Low/High strings at the sensor limits. Libre’s export from LibreView gives separate columns for historic and scan glucose, so filter on record type before computing anything. The first cleaning steps we always take:

import pandas as pd, numpy as np
df = pd.read_csv("clarity.csv", parse_dates=["Timestamp (YYYY-MM-DDThh:mm:ss)"])
df = df[df["Event Type"] == "EGV"].copy()
g = pd.to_numeric(df["Glucose Value (mg/dL)"].replace({"Low": 39, "High": 401}),
                  errors="coerce")
df["glucose"] = g
df = df.set_index("Timestamp (YYYY-MM-DDThh:mm:ss)").sort_index()
# drop the first 12 h of each sensor session: warm-up drift
cv = df["glucose"].std() / df["glucose"].mean() * 100
titr = ((df["glucose"] >= 70) & (df["glucose"] <= 120)).mean() * 100

Discard the first twelve to twenty-four hours of each sensor wear. Interstitial sensors read low during warm-up and are noisiest then, and there is a physiological lag of roughly five to fifteen minutes between plasma and interstitial glucose that widens when glucose is moving fast.56 That lag is why a sensor reading during a post-meal rise understates the plasma peak and why the same sensor overstates glucose on the way down. Compression lows, the sharp artifactual drops that appear when you sleep on the sensor, are the most common artifact people misread as nocturnal hypoglycemia. They look like a fast drop to the 40s and a fast recovery with no intervening plateau, and they are worth flagging and excluding before you compute overnight statistics.

Two more calibration points. Sensor accuracy is specified against a laboratory reference across a wide range, and disagreement of 10–15% at any single point is expected rather than a malfunction.7 And the metrics developed for CGM were largely validated in people with diabetes, where the analytic questions are hypoglycemia detection and time in a 70–180 range.89 Applied to someone whose glucose rarely leaves 80–140, those thresholds are uninformative, which is why we use time in the 70–120 range and CV instead.

For a fasting glucose of 106, the specific things to look for in fourteen days of CGM are the nocturnal profile and the dawn trajectory. If your overnight nadir sits near 95–105 and rises through the last three hours of sleep, you are looking at a dawn pattern driven by cortisol and growth hormone plus hepatic glucose output, which is common and shows up as an elevated fasting value in the lab. If your overnight glucose sits at 85 and your single fasting draw read 106, suspect the draw. If post-meal excursions routinely exceed 160 and take more than two hours to return to baseline, the informative next test is the OGTT with intermediate time points, because that pattern points toward disposal rather than fasting output.

What we would not do

We would not chase a single value down with acute measures. The question “how to bring your glucose level down quickly” belongs to people managing diagnosed diabetes with prescribed medication, and any answer involving drugs or doses is a clinician’s decision. We would also not treat one number as a diagnosis in either direction. A fasting 85 with two-hour post-load values of 180 is not reassuring, and a 106 in someone with a GCK variant is not alarming. The value of building a wider measurement set is that it moves you from a threshold you happen to straddle to a description of which part of glucose handling is doing what in you, and that description is what a physician can act on.

Oak builds longitudinal molecular profiles of individuals: whole-genome sequencing, RNA sequencing, proteomics, blood biomarkers, and continuous glucose data, integrated into one model of you. Build your profile.

Footnotes

  1. Anthony Sallar, Samuel Dagogo-Jack. Regression from prediabetes to normal glucose regulation: State of the science. Experimental Biology and Medicine, 2020. https://doi.org/10.1177/1535370220915644 ↩

  2. Timothy S. Bailey, George. Grunberger, Bruce W. Bode, et al. American Association Of Clinical Endocrinologists And American College Of Endocrinology 2016 Outpatient Glucose Monitoring Consensus Statement. Endocrine Practice, 2016. https://doi.org/10.4158/ep151124.cs ↩

  3. Ramzi Ajjan, David Slattery, Eugene Wright. Continuous Glucose Monitoring: A Brief Review for Primary Care Practitioners. Advances in Therapy, 2019. https://doi.org/10.1007/s12325-019-0870-x ↩

  4. Roger Mazze, Yariv Yogev, Oded Langer. Measuring glucose exposure and variability using continuous glucose monitoring in normal and abnormal glucose metabolism in pregnancy. The Journal of Maternal-Fetal & Neonatal Medicine, 2012. https://doi.org/10.3109/14767058.2012.670413 ↩

  5. D. Barry Keenan, John J. Mastrototaro, Gayane Voskanyan, et al. Delays in Minimally Invasive Continuous Glucose Monitoring Devices: A Review of Current Technology. Journal of Diabetes Science and Technology, 2009. https://doi.org/10.1177/193229680900300528 ↩

  6. Ahmed El-Laboudi, Nick S. Oliver, Anthony Cass, et al. Use of Microneedle Array Devices for Continuous Glucose Monitoring: A Review. Diabetes Technology & Therapeutics, 2013. https://doi.org/10.1089/dia.2012.0188 ↩

  7. Parizad Avari, Alistair Lumb, Daniel Flanagan, et al. Continuous Glucose Monitoring Within Hospital: A Scoping Review and Summary of Guidelines From the Joint British Diabetes Societies for Inpatient Care. Journal of Diabetes Science and Technology, 2022. https://doi.org/10.1177/19322968221137338 ↩

  8. Cornelis A J van Beers, J Hans DeVries, Susanne J Kleijer, et al. Continuous glucose monitoring for patients with type 1 diabetes and impaired awareness of hypoglycaemia (IN CONTROL): a randomised, open-label, crossover trial. The Lancet Diabetes & Endocrinology, 2016. https://doi.org/10.1016/s2213-8587(16)30193-0 ↩

  9. Thomas C. Blevins, Bruce W. Bode, Satish K. Garg, et al. Statement by the American Association of Clinical Endocrinologists Consensus Panel on Continuous Glucose Monitoring. Endocrine Practice, 2010. https://doi.org/10.4158/ep.16.5.730 ↩