A Fasting Glucose of 85 mg/dL Is Normal, and One Reading Tells You Almost Nothing

A fasting plasma glucose of 85 mg/dL is normal by every clinical standard in use. The American Diabetes Association defines normal fasting glucose as below 100 mg/dL, impaired fasting glucose as 100 to 125 mg/dL, and diabetes as 126 mg/dL or higher on two separate occasions.¹ So 85 is not borderline, not “pre-prediabetic,” and not a number that needs fixing. The number is also not informative. A single fasting value is one sample from a distribution that, in people without diabetes, is wider than most readers expect, and the width of that distribution is the part worth measuring.

What a fasting value of 85 measures

“Fasting” in the diagnostic sense means no caloric intake for at least eight hours.¹ Under those conditions your plasma glucose reflects hepatic glucose output (glycogenolysis plus gluconeogenesis) balanced against basal insulin-mediated and insulin-independent uptake. It is a snapshot of the overnight steady state, taken at whatever moment the phlebotomist drew blood.

Several things move that snapshot by amounts comparable to the difference between 85 and 96. The dawn phenomenon, a cortisol- and growth-hormone-driven rise in hepatic glucose output in the pre-waking hours, means a 6:00 a.m. draw and an 8:30 a.m. draw are not interchangeable. Poor sleep the night before, a late dinner that extended past the eight-hour window, an argument in the parking lot, and a hard workout the previous evening all shift fasting glucose. Pre-analytical handling matters too: glucose in a plain serum tube continues to be consumed by red cells at roughly 5 to 7 mg/dL per hour at room temperature, so a sample that sat before centrifugation reads low. Sodium fluoride (gray-top) tubes inhibit glycolysis but do so slowly, over an hour or more. If you want to compare your own values over time, ask what tube type and processing delay your lab uses and hold it constant.

Intrapersonal variability is the number that matters

The most useful recent work on this question used continuous glucose monitoring (CGM) in a large cohort without diabetes and asked a simple thing: if you measure someone’s fasting glucose on many different days, how much does it move? The answer was substantial. In that study, intrapersonal variability in CGM-derived fasting glucose was large enough that a single measurement frequently misclassified individuals relative to their own multi-day average. A meaningful fraction of people whose average fasting glucose fell in the prediabetic range had individual days that looked entirely normal, and vice versa.² The practical implication is that the distance between 85 and 96 is often within one person’s own day-to-day spread.

This reframes the question people are asking. “Is 85 good and 96 bad?” presumes each number is a stable property of the person. Treat your fasting glucose as a random variable with a mean and a standard deviation, and you get two parameters instead of one. Someone averaging 88 with a standard deviation of 4 mg/dL and someone averaging 88 with a standard deviation of 14 mg/dL are metabolically different, and only the second is worth a conversation with a physician about further workup.

The 72 to 85 “optimal range” claim

A widely circulated claim holds that fasting glucose above 85 mg/dL carries elevated risk and that 72 to 85 is the optimal target. The epidemiology behind this is real in outline: in large observational cohorts, cardiovascular and diabetes incidence rise monotonically across the normoglycemic range rather than stepping up at 100 mg/dL. But the inference from that to an individual target has two problems. First, the hazard ratios across the 85-to-99 band are small and heavily confounded by adiposity, fitness, and the metabolic syndrome cluster, so the glucose value is partly acting as a marker for things it does not cause. Second, the measurement precision required to place yourself inside a 13 mg/dL window is greater than a single annual draw provides, given the variability described above.²

Our view: do not chase a fasting number inside the normal range. Chase the shape of your postprandial curves and your fasting variability, both of which are measurable with much better signal-to-noise than a point estimate of the mean.

How to measure this yourself

If you have a CGM, the fasting estimate you want is not a single morning reading. Extract the raw sensor stream and compute a per-day fasting value from a fixed window, then look at the distribution across two weeks.

CGM data typically exports as CSV at five-minute resolution (288 readings per day). A workable definition is the mean of the 04:00–06:00 interval, which avoids both the late-evening meal tail and most of the dawn rise:

import pandas as pd

df = pd.read_csv("cgm.csv", parse_dates=["timestamp"])
df = df.set_index("timestamp").sort_index()
window = df.between_time("04:00", "06:00")["glucose_mg_dl"]
daily = window.resample("1D").agg(["mean", "count"])
daily = daily[daily["count"] >= 20]          # drop days with sensor gaps
print(daily["mean"].describe())
print("CV%%:", 100 * daily["mean"].std() / daily["mean"].mean())

Two cautions on interpreting the output. CGM measures interstitial fluid glucose, not plasma, and the sensor sits behind a physiological lag of roughly 5 to 10 minutes plus an algorithmic smoothing filter, so overnight flat periods are the regime where CGM and venous plasma agree best and rapid excursions are where they diverge most.³ Factory-calibrated sensors also carry a per-sensor bias: the same person wearing two sensors simultaneously commonly sees a 5 to 15 mg/dL offset between them.⁴ Compare sensors to themselves over their wear period, and anchor the whole series with one venous fasting draw taken the same morning if you want absolute accuracy.

The second useful output is the postprandial response, which is where individual differences are largest and most actionable. For each labeled meal, compute peak glucose and time to peak. Also compute incremental area under the curve over three hours, and the time to return within 15 mg/dL of the pre-meal baseline. A standardized challenge repeated across weeks (same food, same time, same prior-day activity) gives you a comparable test. Fifty grams of glucose in solution is the classic stimulus, but a fixed real meal is more representative of what you eat.

When this stops being a data exercise

Three patterns move you from self-measurement to a clinician’s office. An average fasting glucose above 100 mg/dL across a two-week CGM window, or a venous fasting glucose of 126 mg/dL or higher, meets diagnostic thresholds and needs confirmatory laboratory testing rather than sensor data.¹ Repeated sensor readings below 60 mg/dL in someone not taking glucose-lowering medication are unusual and merit evaluation. And pregnancy changes the targets entirely: glycemic goals in pregnancy are considerably tighter than the non-pregnant normal range, and management is specialized.⁵ Do not self-interpret glucose data in pregnancy.

Questions people also ask

Is 85 bad for blood sugar? No. It sits in the middle of the normal fasting range and is below the 100 mg/dL threshold for impaired fasting glucose.¹ It is a good number and it does not need intervention.

What should my blood sugar be after 8 hours of fasting? Below 100 mg/dL is normal, 100 to 125 mg/dL is impaired fasting glucose, and 126 mg/dL or above on two occasions indicates diabetes.¹ Eight hours is the minimum fast for the test to be interpretable.

What is my A1c if my glucose is 85 or 86? You cannot convert a single fasting glucose to HbA1c. The published regressions map mean glucose over roughly three months to A1c, and a fasting point estimate is a poor proxy for mean glucose because it omits every postprandial excursion. A CGM-derived mean over 14 days is a much better input, though even that relationship has person-to-person scatter driven by red cell lifespan and glycation rate.

Is a 110 fasting sugar level normal? It falls in the impaired fasting glucose range of 100 to 125 mg/dL.¹ Given the day-to-day variability described above, one value of 110 warrants a repeat measurement rather than a conclusion, and the interpretation belongs with a physician.

Is 86 too low? No. Symptomatic hypoglycemia in people not on glucose-lowering medication generally occurs well below 70 mg/dL, and healthy overnight troughs in the 60s are common on CGM.

Is 96 non-fasting meaningful? A value of 96 two hours after eating is a well-controlled response, since the postprandial reference is below 140 mg/dL. Non-fasting values are not diagnostic on their own.

Footnotes

1. American Diabetes Association. 2. Classification and Diagnosis of Diabetes: Standards of Medical Care in Diabetes—2018. Diabetes Care, 2017. https://doi.org/10.2337/dc18-s002 ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶

2. Smadar Shilo, Ayya Keshet, Hagai Rossman, et al. Continuous glucose monitoring and intrapersonal variability in fasting glucose. Nature Medicine, 2024. https://doi.org/10.1038/s41591-024-02908-9 ↩ ↩²

3. David C. Klonoff, David Ahn, Andjela Drincic. Continuous glucose monitoring: A review of the technology and clinical use. Diabetes Research and Clinical Practice, 2017. https://doi.org/10.1016/j.diabres.2017.08.005 ↩

4. Sandeep Vashist. Continuous Glucose Monitoring Systems: A Review. Diagnostics, 2013. https://doi.org/10.3390/diagnostics3040385 ↩

5. Anastasia-Stefania Alexopoulos, Rachel Blair, Anne L. Peters. Management of Preexisting Diabetes in Pregnancy: A Review. Obstetrical & Gynecological Survey, 2019. https://doi.org/10.1097/ogx.0000000000000726 ↩