Why Fasting Glucose Runs High When the Rest of the Day Looks Fine

If your glucose is high after eight, sixteen, or forty-eight hours without food, the food is not the cause. During a fast your liver is the sole supplier of circulating glucose, first from glycogen breakdown and then increasingly from gluconeogenesis, and the rate at which it supplies that glucose is set by hormones rather than by your last meal. Between roughly 03:00 and 08:00, growth hormone and cortisol rise. Hepatic glucose output rises with them, and glucose climbs even though nothing has entered your gut. That is the dawn phenomenon. A fasting value that sits above your daytime baseline while post-meal excursions look unremarkable is the classic signature of hepatic glucose output that insulin is not fully restraining. The pattern is a question for a physician with a lab panel. A sensor trace is not a basis for self-diagnosis.

What drives glucose during a fast

In the first twelve to sixteen hours without food, most of the glucose in your blood came from liver glycogen. As glycogen depletes, gluconeogenesis from lactate, glycerol, and amino acids takes over, and free fatty acids released from adipose tissue both fuel that process and reduce insulin’s ability to suppress it. The result is that glucose settles onto a plateau on a long fast, and in many people that plateau drifts upward across days two and three rather than downward. Glucose does not fall to zero.

Layered on top is the circadian surge. Growth hormone pulses in early sleep, cortisol peaks near waking, and both raise hepatic glucose production over a window of several hours. Two people with identical daytime curves can differ by 20 to 30 mg/dL at 07:00 purely on the amplitude of that surge. Sleep debt, alcohol the previous evening, a late high-fat meal, and hard exercise the day before all shift it, which is why a single morning fingerstick tells you far less than two weeks of overnight traces.

Rule out the measurement before you explain the number

A large share of surprising fasting numbers are artifacts, and it is cheap to check. Continuous glucose monitors measure glucose in interstitial fluid, not blood, and report through a factory or user calibration with a mean absolute relative difference in the single digits to low teens depending on the system and the glucose range ¹². That means a displayed 108 mg/dL is compatible with a true value in the high 90s or high 110s. Overnight the error is not random: lying on the sensor compresses tissue and produces false lows that rebound sharply when you roll off, which reads on the chart as a spike at 04:00 that never happened.

Three practical filters. First, discard the first 12 to 24 hours after insertion, where sensors commonly read low and drift. Second, treat any overnight excursion that appears as a sharp V or inverted V within 15 to 20 minutes as suspect until it repeats on a different night in a different sleep position. Third, if a fasting number matters, confirm it with a venous plasma glucose drawn in a lab. The sample should be collected in a sodium fluoride tube and spun promptly, because glucose in a plain tube falls roughly 5 to 7 percent per hour from red cell glycolysis at room temperature.

Pulling your own data and computing the right numbers

Both major systems export clean CSVs. Dexcom Clarity gives you Timestamp (YYYY-MM-DDThh:mm:ss), Event Type, and Glucose Value (mg/dL) at 5-minute resolution. Values below 40 and above 400 appear as the strings Low and High and will break a naive astype(float). Abbott’s Libre export uses Device Timestamp in local time with a Record Type column where 0 is the 15-minute historic reading and 1 is a scan, and you almost always want to keep only Record Type == 0 so that manual scans do not double-weight the daytime.

A minimal pipeline in pandas:

df = pd.read_csv("clarity.csv", skiprows=1)
df = df[df["Event Type"] == "EGV"].copy()
df["t"] = pd.to_datetime(df["Timestamp (YYYY-MM-DDThh:mm:ss)"])
df["g"] = pd.to_numeric(df["Glucose Value (mg/dL)"]
                        .replace({"Low": 39, "High": 401}), errors="coerce")
df = df.set_index("t").resample("5min").mean().interpolate(limit=3)

From there, four quantities describe an overnight far better than one fasting reading does. The nocturnal nadir is the minimum between 00:00 and 05:00. The dawn amplitude is the waking value minus that nadir, computed per night and then summarized as a median across nights rather than an average, because one compression artifact will wreck the mean. Overnight coefficient of variation is the standard deviation divided by the mean across 00:00 to 06:00. It should be low in a stable fast, and a value above about 15 percent usually points at sensor noise or disrupted sleep rather than metabolism. Finally, the median 06:00 to 08:00 value across at least ten nights is the number worth comparing to a lab draw. Plot each night as a separate faint line on shared axes before you summarize anything. The per-night shapes tell you whether you are looking at a consistent dawn rise, a post-alcohol reactive pattern, or intermittent noise.

Fasting high, post-meal normal

This combination is common and it is informative. Fasting glucose is largely a readout of hepatic glucose production and how well basal insulin suppresses it. Post-meal glucose is largely a readout of how quickly skeletal muscle disposes of an incoming load, alongside first-phase insulin secretion. The two can diverge in either direction, which is why a normal HbA1c and a normal fasting plasma glucose do not rule out abnormal glucose handling elsewhere in the day. Continuous monitoring in people without a diabetes diagnosis detects dysglycemia that both HbA1c and fasting glucose miss ³. Conversely, reference CGM data from healthy adults show that even people with entirely normal labs spend brief periods above 140 mg/dL after meals, so isolated peaks are not by themselves meaningful ⁴.

The practical consequence is that you should not try to reason about your metabolism from a single measurement class. HbA1c integrates roughly three months of glucose but is distorted by anything that changes red cell lifespan, including iron deficiency and hemoglobin variants, which is one reason a genome file worth checking for HBB variants is useful context. A fasting plasma glucose is one point on one morning. Fourteen days of CGM plus a lab fasting insulin and glucose drawn together give you the overnight dynamics and the hormonal context at once, and interpreting them together is a clinician’s job.

Multi-day and religious fasts

Extended fasting does not reliably lower glucose, and in people on glucose-lowering medication it produces both directions of excursion. CGM during Ramadan in people with diabetes shows substantial daytime hypoglycemia alongside post-iftar hyperglycemia, patterns that self-monitored fingersticks systematically under-detect ⁵. Similar work in high-risk patients with chronic kidney disease and in gestational diabetes found that continuous monitoring exposed excursions that intermittent testing missed entirely ⁶⁷. If you take any medication that lowers glucose, planning an extended fast without medical supervision is unsafe, and the monitoring question is secondary to that.

When the answer is a clinician and a script will not do

Persistent fasting values above the standard laboratory cut points, a rising trend across months, symptoms such as unexplained weight loss or heavy thirst, or any fasting number that surprises you on a repeat lab draw all belong with a physician. The value of doing the analysis yourself is that you arrive with two weeks of per-night traces, a median dawn amplitude, and a confirmed venous draw instead of a single alarming reading. Real-time monitoring has been shown to change glycemic outcomes in type 2 diabetes even without medication changes, largely through the feedback it provides ⁸, and the same feedback loop is what makes your own export worth reading carefully.

Questions people also ask

How long does it take for blood sugar to drop after fasting? After a mixed meal most people return to their baseline within two to three hours, and the CGM curve is back near pre-meal level by hour three ⁴. Beyond that point, glucose is set by hepatic output and stops falling, which is why hour 16 of a fast often reads higher than hour 4.

Why is my blood sugar still high while fasting? Because your liver is producing it. Glycogenolysis and gluconeogenesis continue throughout a fast, and the early-morning rise in growth hormone and cortisol increases that output regardless of when you last ate.

Why is my fasting blood sugar high but normal after eating? Fasting values reflect basal insulin’s restraint of hepatic glucose production, while post-meal values reflect muscle glucose uptake and the acute insulin response. Those two systems can be affected differently, and the divergence is worth showing to a physician alongside a paired fasting insulin and glucose.

Why does glucose spike after fasting? Two separate causes. A genuine physiological rise near waking from counterregulatory hormones, or a sensor compression artifact from lying on the device, which appears as a sharp dip and rebound over 15 to 20 minutes. Checking the shape and the sleep position distinguishes them.

Does a CGM replace a lab test? No. Sensors carry meaningful error against plasma glucose ¹², and systematic reviews of real-time monitoring treat it as a complement to laboratory measurement rather than a substitute ⁹. Use CGM for dynamics and a venous draw for the absolute number.

Footnotes

1. Sandeep Vashist. Continuous Glucose Monitoring Systems: A Review. Diagnostics, 2013. https://doi.org/10.3390/diagnostics3040385 ↩ ↩²

2. 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 ↩ ↩²

3. Santiago Rodriguez-Segade, Javier Rodriguez, Félix Camiña, et al. Continuous glucose monitoring is more sensitive than HbA1c and fasting glucose in detecting dysglycaemia in a Spanish population without diabetes. Diabetes Research and Clinical Practice, 2018. https://doi.org/10.1016/j.diabres.2018.05.026 ↩

4. Paul R.E. Jarvis, Jessica L. Cardin, Pamela M. Nisevich-Bede, et al. Continuous glucose monitoring in a healthy population: understanding the post-prandial glycemic response in individuals without diabetes mellitus. Metabolism, 2023. https://doi.org/10.1016/j.metabol.2023.155640 ↩ ↩²

5. N. Lessan, Z. Hannoun, H. Hasan, et al. Glucose excursions and glycaemic control during Ramadan fasting in diabetic patients: Insights from continuous glucose monitoring (CGM). Diabetes & Metabolism, 2015. https://doi.org/10.1016/j.diabet.2014.11.004 ↩

6. Fatheya Alawadi, Fauzia Rashid, Alaaeldin Bashier, et al. The use of Free Style Libre Continues Glucose Monitoring (FSL-CGM) to monitor the impact of Ramadan fasting on glycemic changes and kidney function in high-risk patients with diabetes and chronic kidney disease stage 3 under optimal diabetes care. Diabetes Research and Clinical Practice, 2019. https://doi.org/10.1016/j.diabres.2019.03.015 ↩

7. Bachar Afandi, Mohamed Hassanein, Sana Roubi, et al. The value of Continuous Glucose Monitoring and Self-Monitoring of Blood Glucose in patients with Gestational Diabetes Mellitus during Ramadan fasting. Diabetes Research and Clinical Practice, 2019. https://doi.org/10.1016/j.diabres.2019.01.036 ↩

8. Robert A. Vigersky, Stephanie J. Fonda, Mary Chellappa, et al. Short- and Long-Term Effects of Real-Time Continuous Glucose Monitoring in Patients With Type 2 Diabetes. Diabetes Care, 2011. https://doi.org/10.2337/dc11-1438 ↩

9. L. B. E. A. Hoeks, W. L. Greven, H. W. de Valk. Real-time continuous glucose monitoring system for treatment of diabetes: a systematic review. Diabetic Medicine, 2011. https://doi.org/10.1111/j.1464-5491.2010.03177.x ↩