What a Morning Blood Sugar of 115 Tells You

A morning fasting glucose of 115 mg/dL means your plasma glucose was 115 mg/dL at the moment the needle went in. It does not, by itself, tell you your average glucose, your insulin sensitivity, or your trajectory. The 100 mg/dL threshold you will find on every clinic page is a population cutoff drawn on a continuous distribution, and a single draw sits inside a measurement process with enough variance to move you across it. The useful response is to measure more and measure differently, then take the fuller picture to a clinician who can interpret it in context.

The variance in a single fasting draw

Fasting plasma glucose has substantial within-person biological variation. The same person, sampled on consecutive mornings under identical fasting conditions, will produce readings that differ by several mg/dL in either direction. Coefficients of variation around 5 to 7 percent are typical, which puts a one-sigma band of roughly 6 to 8 mg/dL around a value of 115. Two draws a week apart reading 108 and 122 are compatible with a stable underlying state.

On top of biological variation sits pre-analytical error, and this one is larger than most people expect. Glucose in a whole-blood tube continues to be consumed by red and white cells after the draw. In a plain serum or lithium heparin tube left at room temperature, glycolysis lowers the measured value by roughly 5 to 7 percent per hour. Sodium fluoride gray-top tubes slow this but do not stop it quickly. So the direction of pre-analytical error is almost always downward. A sample that sat before centrifugation reads low, which means a reported 115 could correspond to a higher true value at the time of the draw.

Two practical consequences. First, never revise your understanding of your metabolism on one fasting glucose. Second, when you do draw, ask whether the lab uses fluoride-citrate tubes and how long the sample waited. That single detail moves your number more than most lifestyle changes would.

Why “106 not fasting” is a different question

A non-fasting glucose of 106 mg/dL carries almost no diagnostic weight, because the reference intervals printed on your lab report assume a fast of 8 or more hours. Post-meal glucose in metabolically healthy adults routinely rises into the 120s and 140s and returns to baseline within two hours. A random 106 taken three hours after lunch is unremarkable. A random 106 taken 30 minutes after a large bowl of rice is, if anything, a sign of brisk glucose disposal.

The practical error here is comparing a postprandial value against a fasting range and concluding something is wrong. If you want a number that is comparable to the reference range, fast properly for 10 to 12 hours. That means water only, with no coffee and no nicotine. If you want to understand your postprandial physiology, you need a time series rather than a point.

The four measurements that resolve the ambiguity

We would not repeat a single fasting glucose and hope for a cleaner number. We would run four measurements that answer different questions.

HbA1c gives you a weighted average of glucose over roughly the preceding 8 to 12 weeks, with the most recent month dominating. It is reported as a percentage under NGSP standardization and in mmol/mol under IFCC. It is a red cell measurement, so it is biased by anything that changes erythrocyte lifespan. Iron deficiency and low reticulocyte turnover push A1c up. Hemolysis, recent blood donation, and some hemoglobin variants push it down. Always read A1c next to a CBC, and specifically next to hemoglobin, MCV, and RDW.

Fasting insulin, drawn from the same tube as the fasting glucose, tells you what it cost your body to hold that glucose value. A fasting glucose of 115 with insulin of 4 µIU/mL describes a different system than 115 with insulin of 22 µIU/mL. From the pair you can compute HOMA-IR as (glucose in mg/dL times insulin in µIU/mL) divided by 405. Treat the absolute value with suspicion, because insulin assays are poorly harmonized across platforms, and use it mainly to track yourself over time on one assay.

Continuous glucose monitoring gives you the shape of the curve rather than a point on it. A sensor sampling every 5 minutes produces about 288 values per day, which is roughly 2,000 measurements over a week. That is the difference between one draw and a distribution. CGM in people without diabetes has moved from curiosity to a reasonable research and self-measurement tool, though the interpretive frameworks are still being worked out ¹.

An oral glucose tolerance test, 75 g of glucose with draws at 0, 30, 60, 90 and 120 minutes, remains the most informative single provocation. The 30-minute and 60-minute values carry information that fasting glucose and A1c both miss. Order it with insulin at each timepoint if you can, since the insulin curve shape is the more interesting half.

What the CGM trace tells you about morning highs

If your fasting glucose is 115, the most common CGM finding is an overnight floor around 95 to 105 with a pre-waking rise. Cortisol and growth hormone secretion in the hours before waking raise hepatic glucose output, and glucose climbs 10 to 20 mg/dL without any food involved. Waking at 115 after a night spent at 100 is a different pattern from sitting at 115 all night, and only the trace distinguishes them.

Look for four things in a two-week trace. The overnight nadir and its stability across nights. The magnitude and timing of the pre-waking rise. Peak height and time-to-return-to-baseline after your repeated meals. And the total variability, expressed as coefficient of variation, which for people without diabetes typically falls below about 20 percent.

Read the trace with its limitations in mind. CGM measures interstitial fluid, not plasma, and lags plasma by roughly 5 to 15 minutes during rapid change. Sensors are factory-calibrated with mean absolute relative difference usually in the 8 to 12 percent range, and accuracy is worst in the first 24 hours after insertion and at the low end of the range. Compression artifacts from sleeping on the sensor produce false nocturnal lows that resolve when you roll over. Exercise adds another layer, since sustained aerobic work changes both glucose kinetics and sensor behavior in ways that complicate interpretation ². Sport-specific work in runners has shown how much the trace can move under load ³.

Meal timing is one of the few variables with clean interventional evidence. Early time-restricted eating, with the feeding window ending in the late afternoon, improved 24-hour glucose levels in a controlled crossover trial in men with prediabetes ⁴. That study used the same instrument you would be using, which makes it a reasonable template for structuring your own n-of-1 comparison.

How we would run the analysis

Export the raw CGM data rather than working inside the vendor app. Dexcom Clarity and Abbott LibreView both export CSV with a timestamp and a glucose value, though the column names and datetime formats differ between them and between firmware versions. Parse with pandas.read_csv, then immediately coerce timestamps with pd.to_datetime(..., errors="coerce") and check how many rows failed. Set the index and resample to a regular 5-minute grid with df.resample("5min").mean(), since real exports have gaps from sensor warmups and dropouts.

Compute the standard metrics yourself so you know what they mean. Mean, standard deviation, and CV over the full record. Time in range as the fraction of readings between 70 and 140 mg/dL. Overnight nadir as the minimum over a fixed window, say 02:00 to 05:00, computed per night so you see the spread. For meal responses, take the incremental area under the curve over 120 minutes from meal start, using trapezoidal integration with the pre-meal value as baseline.

Then structure an actual experiment. Pick one standardized test meal, eat it at the same time of day on at least six separate days, and compare against six days of a single alternative. Six repeats of each is a reasonable floor, because day-to-day variability in postprandial response within one person is large enough to make three comparisons uninformative. Keep sleep and prior-day exercise in your notes, since both shift next-morning glucose.

Finally, be clear about the boundary. All of this is measurement and interpretation. A fasting glucose of 115, an A1c in the prediabetic range, or a CGM trace with frequent excursions above 180 mg/dL are findings to bring to a physician. The physician can order confirmatory testing and decide what, if anything, should be done. We do not recommend treatments, and no blog post can substitute for that conversation. If you are pregnant, the thresholds and the stakes are different, and this belongs with your obstetric team rather than in a self-analysis notebook ⁵.

Questions people also ask

Is a blood sugar of 115 in the morning prediabetes?

A fasting value of 100 to 125 mg/dL falls in the range labeled impaired fasting glucose, but a diagnosis rests on repeated testing rather than one draw. Get a second fasting glucose plus an A1c, ideally with fasting insulin from the same tube, and let a clinician interpret the set.

Is a glucose of 106 while not fasting a problem?

A non-fasting 106 mg/dL tells you very little, because the reference range on your report assumes an 8-hour fast. Normal postprandial glucose commonly reaches the 120s or 140s and returns toward baseline within two hours.

Why is my blood sugar highest in the morning?

Cortisol and growth hormone rise in the hours before waking and increase hepatic glucose output, which lifts glucose without any food. This produces a pre-waking climb of roughly 10 to 20 mg/dL in many people, and a CGM trace will show whether that is your pattern.

Does coffee before the draw affect fasting glucose?

Caffeine can raise glucose modestly and acutely reduce insulin sensitivity, so black coffee breaks the assumptions of a fasting draw. Water only for 10 to 12 hours before the test.

How long should I wear a CGM to learn something useful?

Two weeks is a reasonable minimum, since 14 days at 5-minute sampling gives about 4,000 readings and enough repeated days to separate signal from noise. If you are testing a specific meal or timing change, plan on at least six repeats of each condition.

Footnotes

1. David C. Klonoff, Kevin T. Nguyen, Nicole Y. Xu, et al. Use of Continuous Glucose Monitors by People Without Diabetes: An Idea Whose Time Has Come?. Journal of Diabetes Science and Technology, 2022. https://doi.org/10.1177/19322968221110830 ↩

2. Mikael Flockhart, Filip J. Larsen. Continuous Glucose Monitoring in Endurance Athletes: Interpretation and Relevance of Measurements for Improving Performance and Health. Sports Medicine, 2023. https://doi.org/10.1007/s40279-023-01910-4 ↩

3. E. Cauza, U. Hanusch-Enserer, B. Strasser, et al. Continuous Glucose Monitoring in Diabetic Long Distance Runners. International Journal of Sports Medicine, 2005. https://doi.org/10.1055/s-2004-830561 ↩

4. Humaira Jamshed, Robbie Beyl, Deborah Della Manna, et al. Early Time-Restricted Feeding Improves 24-Hour Glucose Levels and Affects Markers of the Circadian Clock, Aging, and Autophagy in Humans. Nutrients, 2019. https://doi.org/10.3390/nu11061234 ↩

5. Dorota Bomba-Opoń, Katarzyna Cypryk, Krzysztof Czajkowski, et al. Continuous Glucose Monitoring in Gestational Diabetes Mellitus: Evidence, Clinical Benefits and Remaining Challenges. Diabetes Therapy, 2026. https://doi.org/10.1007/s13300-026-01906-z ↩