LabNorms Population Percentiles

Fasting Glucose

Fasting plasma glucose measures blood sugar after an overnight fast and rises steadily with age, with the upper tail widening substantially in older adults due to the prevalence of undiagnosed and diagnosed diabetes in the general population.

Unit: mg/dL · 12 groups · age and sex · 1 source

Filed under panels: Diabetes Panel , Metabolic Panel · topics: Metabolic

Fasting Glucose by age and sex

Unit:
Median and middle 90% (P5 to P95), NHANES 2017-March 2020 Pre-Pandemic. Select a cell for the full distribution. US civilian non-institutionalized adults aged 20 and over, not screened for health. Measured after a fast of at least 8 hours. Population percentiles, not clinical reference intervals.
Age (years) Males (mg/dL)Females (mg/dL)
20-29 median 99 middle 90% 87 to 111 median 94 middle 90% 83 to 114
30-39 median 101 middle 90% 87 to 127.4 median 97 middle 90% 86 to 131
40-49 median 104 middle 90% 91 to 171 median 100 middle 90% 86 to 149
50-59 median 107 middle 90% 91 to 278.5 median 103 middle 90% 87 to 158.6
60-69 median 112 middle 90% 94 to 200 median 106 middle 90% 89 to 165.3
70+ median 113 middle 90% 93 to 195.9 median 105 middle 90% 90 to 152.3

Download Fasting Glucose percentiles (CSV) Values at full stored precision; tables on this site may round for display. Licensed under CC BY 4.0.

Population Distribution

Fasting plasma glucose (FPG) is measured after a minimum 8-hour fast. These percentiles are derived from the full US population and include individuals with prediabetes and diabetes. The distribution is roughly symmetric in younger adults but becomes increasingly right-skewed with age as metabolic dysfunction becomes more common. Note: 1 mg/dL is approximately 0.0555 mmol/L. Corrected October 2026: the previous version of these figures included participants who did not meet the fasting requirement of 8 to under 24 hours. Removing them lowered each group's participant count by 7 to 27 and changed 10 of the 84 percentiles.

How it was measured

Sample
Plasma
Method
Hexokinase enzymatic assay
Instrument
Roche/Hitachi Cobas c311
Laboratory
University of Missouri Diabetes Diagnostic Laboratory
Fasting
Fasted 8 to under 24 hours before the blood draw (NHANES fasting subsample) (NHANES fasting questionnaire documentation, P_FASTQX)

Results from another laboratory can differ from these because of differences in method, calibration, and sample handling.

Sources: NHANES 2017-March 2020 laboratory documentation, P_GLU · Fasting glucose procedure manual, 2017-2018 · Fasting glucose procedure manual, 2019-2020

Frequently Asked Questions

Why do these percentiles include people with diabetes?

These are population percentiles, not clinical reference ranges. They describe where values fall across the full US population, including people with prediabetes and diabetes. Clinical reference ranges (from labs like Labcorp or Quest) are derived from screened, presumed-healthy individuals and answer a different question.

Why does the upper end of the distribution rise so sharply with age?

As age increases, a growing proportion of the population has impaired glucose regulation, prediabetes, or diabetes. These individuals pull the upper tail of the distribution upward, so P75 and especially P95 increase substantially in older age groups even though typical (median) values rise only modestly.

How do I convert mg/dL to mmol/L?

Divide by 18.02 (or multiply by 0.0555). For example, 100 mg/dL is approximately 5.6 mmol/L. The mmol/L unit is used in most countries outside the United States.

Data Sources

Cite this data

LabNorms. Fasting Glucose population percentiles by age and sex. Data: NHANES 2017-March 2020 Pre-Pandemic. https://labnorms.com/analytes/fasting-glucose/. Licensed under CC BY 4.0. Accessed [date].

Download Fasting Glucose percentiles (CSV) Values at full stored precision; tables on this site may round for display. Licensed under CC BY 4.0.

Related Analytes

HbA1c

Long-term glycemic marker

Fasting Insulin

Companion fasting glycemic marker

HOMA-IR

Primary input used to derive HOMA-IR