Lab Reference Range Chart

Start with the second column of the first table, because it is the thing every other lab chart leaves out. Two completely different kinds of number get printed in the same typeface on a reference sheet. A REFERENCE INTERVAL is a percentile range measured in a named population on a named assay — it describes where most healthy people fall. A DECISION THRESHOLD is a line a guideline committee drew, usually where risk starts rising, and it is a percentile of nothing. Of the four analytes on this sheet, exactly one is the first kind.

That one is total testosterone: 264–916 ng/dL, the 2.5th to 97.5th percentiles of healthy nonobese men aged 19 to 39, harmonized across four cohorts by cross-calibrating every assay to the CDC reference method. It is a real interval and it comes with three conditions. It is a young man's interval applied unchanged at 70. It assumes a calibrated assay. And being inside it is not a clean bill of health, just as being marginally outside it is not a diagnosis.

The other three are decisions, and the sheet labels them as such. NIDDK publishes the A1C figures in a table headed "Diagnosis", with a footnote requiring a second confirmatory measurement. KDIGO attaches a footnote to its GFR categories saying that G1 and G2 do not meet the criteria for chronic kidney disease at all without separate evidence of kidney damage. The lipid numbers are classification band boundaries from a paper about an estimating equation, and the US and European lists do not even agree with each other. Treating any of them as "your normal range" is the mistake this layout is designed to prevent.

The last table is the one that makes this sheet different. Eight ranges that appear on nearly every men's lab chart are named there and explicitly not printed, with what we found when we went looking: no harmonized interval for calculated free or bioavailable testosterone, no established range for estradiol in men, no sourced population value for total cholesterol or HDL, and a decade-by-decade testosterone staircase whose citation trail runs out at a newsletter. If a figure is missing from this sheet, it is missing on purpose and it is listed.

Finally: read your own result against the range your own laboratory printed on your own report. A reference interval belongs to the assay that produced it. This sheet tells you what kind of number you are looking at and where it came from; it does not replace the interval your lab states, and nothing on it is a diagnosis.

What kind of number each figure is

The column that matters most on any reference chart, and the one no other chart carries. A reference interval describes a measured population. A decision threshold is a line a committee drew. They are not the same thing and they are printed in the same typeface everywhere else.

Four analytes are printed and five are refused, which is roughly the ratio you would get on any honest men's lab sheet. Of the four printed, one is a reference interval and three are decisions. Note in particular that the testosterone row is the only one whose "population" cell names a cohort at all — the other three name a committee.

Total testosterone — the harmonized reference interval

The only figure on this sheet derived as a percentile of a measured population. Percentiles are as the paper publishes them; nmol/L is converted, not separately quoted.

Travison TG, et al. J Clin Endocrinol Metab. 2017;102(4):1161–1173 (PMID 28324103). Its abstract states: "In healthy nonobese men, 19 to 39 years, harmonized 2.5th, 5th, 50th, 95th, and 97.5th percentile values were 264, 303, 531, 852, and 916 ng/dL, respectively." Three things this table cannot be read without. It is a reference interval for MEN AGED 19 TO 39 and is applied unchanged to a man of 70 — a deliberate choice the paper argues in print, not an oversight. It assumes an assay cross-calibrated to the CDC reference method; results from an uncalibrated direct immunoassay are not interchangeable with it. And a percentile is not a treatment threshold: guidelines use 264, 275, 300 and 317 ng/dL for different purposes and none of them moves with age. For the decade-by-decade centiles, and for the decade table we could not source, see /charts/testosterone-levels-by-age-chart.

HbA1c — two thresholds, and what they mean in glucose

NIDDK publishes these as a diagnosis table, not a reference interval. The glucose columns are the ADAG regression, which has real scatter — the published 95% limits are the fourth column and they are wide.

NIDDK's table is headed "Diagnosis" and reads: "Normal below 5.7 percent / Prediabetes 5.7 to 6.4 percent / Diabetes 6.5 percent or above", with the footnote "Any test used to diagnose diabetes requires confirmation with a second measurement, unless there are clear symptoms of diabetes." Those are lines drawn by a guideline, not percentiles of a population, and the word "normal" in that table is a band label rather than a reference interval. It also specifies the assay: "your doctor will send your blood sample taken from a vein to a lab that uses an NGSP-certified method." The glucose columns are Nathan 2008's regression, AG(mg/dl) = 28.7 × A1C − 46.7, R² = 0.84, fitted on 507 subjects and valid over A1C 4–12%. The fourth column is why the point estimate should never be read alone: at 8% the paper's own 95% limits run 147 to 217 mg/dL for a point estimate of 183.

Lipids — band boundaries, and no row for total cholesterol or HDL

Every lipid figure this site could attribute to a named source, converted with the right factor for its molecule. None of these is a target and none is a reference interval.

There is no row for total cholesterol and no row for HDL cholesterol, and that is deliberate rather than an omission: no source this site opened attributes a normal or optimal value for either to a named population, so neither appears. Nor is any row above a target — the US and European classifications are not even the same, cutting at 130 and 160 in one and at 155 in the other. Note also that cholesterol and triglycerides convert with different factors, 0.0259 and 0.0113, because the molecules differ in mass by roughly 2.3 times; using one factor for both is the commonest error on a lipid report. And an LDL-C on a US report is calculated rather than measured, by one of three equations that disagree by up to 12 mg/dL on the same blood — see /tools/ldl-calculator.

eGFR — KDIGO categories, which are not a diagnosis

The staging scheme, in KDIGO's own words. A category describes one number on one day; chronic kidney disease is a diagnosis that requires the reduction to persist beyond three months.

KDIGO 2024, Table 2. Its own footnote reads: "Relative to the young adult level. In the absence of evidence of kidney damage, neither G1 nor G2 fulfills the criteria for CKD." The guideline's threshold is "GFR <60 ml/min per 1.73 m2 (GFR categories G3a-G5) for >3 months to indicate a diagnosis of CKD". eGFR is indexed to 1.73 m² of body surface area, so it is a rate per standard body size rather than a measurement of one man's kidneys, and it assumes creatinine is in steady state — which it is not after a hard training session, on a high-protein diet, on creatine supplementation, or at extremes of muscle mass. Several of those describe this site's readership rather than rare exceptions.

Figures this sheet declines to print

Every range a reader arrives on a men's lab chart expecting, that we could not attribute to a named population measured on a named assay. Named here rather than quietly left out.

Eight refusals against four printed figures. If you have seen a men's lab chart with thirty rows and no citations, this is the difference. Where this site has set out the evidence for a refusal at length, the tool page says so: /tools/testosterone-unit-converter and /tools/bioavailable-testosterone-calculator for calculated free and bioavailable testosterone, /tools/estradiol-unit-converter for estradiol in men, /tools/cholesterol-unit-converter for total cholesterol and HDL, and /charts/testosterone-levels-by-age-chart for the decade table.

Frequently asked questions

What are normal lab values for a man?

Fewer than any chart implies, and the honest list is short. The one men's analyte with a harmonized population reference interval is total testosterone, at 264–916 ng/dL (9.2–31.8 nmol/L) in healthy nonobese men aged 19 to 39. HbA1c, LDL cholesterol and eGFR all have widely used numbers — 5.7% and 6.5%, the 70/100/130/160/190 mg/dL LDL bands, the 90/60/45/30/15 GFR cuts — but those are guideline decisions rather than percentiles of a measured population, and this sheet labels them as such. For total cholesterol, HDL, calculated free testosterone, bioavailable testosterone and estradiol in men, no range is printed here at all, because we could not attribute one to a named population.

What is the difference between a reference range and a normal range?

In careful use there is none — both mean the interval, usually the 2.5th to 97.5th percentile, into which most of a defined healthy population falls. The confusion is with a third thing that gets printed in the same column: a diagnostic threshold. An A1C of 5.7% is not the 2.5th or 97.5th percentile of anything; it is where a guideline decided prediabetes begins. NIDDK prints it in a table headed "Diagnosis" and labels the row below 5.7% "Normal", which is a band name rather than a measured interval. The distinction matters because a percentile tells you how unusual you are and a threshold tells you what someone decided to do about it.

Why does this chart not give a normal free testosterone range?

Because none has been published that could honestly be quoted. Travison 2017 harmonized TOTAL testosterone by cross-calibrating four cohorts to the CDC reference method; no equivalent exists for calculated free testosterone, and there is no CDC standardisation programme for it either. The same is true of calculated bioavailable testosterone. Three other pages on this site say exactly this and mark no row normal, and this sheet does not contradict them. A calculated free testosterone is still useful — it moves when SHBG moves and a total does not — but it should be read against the interval your own laboratory prints for its own equation and assay.

Are testosterone reference ranges different by age?

The one in clinical use is not, and that is a deliberate decision rather than an oversight. 264–916 ng/dL was derived from healthy nonobese men aged 19 to 39 and is applied unchanged to a man of 70; Travison 2017 published age-stratified centiles alongside it and argued the choice both ways in print. Age-specific centiles do exist and are reproduced on /charts/testosterone-levels-by-age-chart. What does not exist, anywhere we could trace, is the decade staircase that appears on most sites — 40–49: 252–916, 50–59: 215–878, 60–69: 196–859 — whose citation chain runs consumer article to laboratory blog to a newsletter and stops.

What does an eGFR of 75 mean?

It is KDIGO category G2, which the guideline words as "Mildly decreased" — and KDIGO's own footnote to that table says "In the absence of evidence of kidney damage, neither G1 nor G2 fulfills the criteria for CKD." So 75 on its own is not chronic kidney disease. The guideline's diagnostic threshold is a GFR below 60 sustained for more than three months, and above 60 a diagnosis additionally needs a marker of damage such as albuminuria, which is a urine test rather than a blood one. eGFR also assumes creatinine is in steady state: creatine supplementation, a high-protein diet, a hard session in the gym and a large muscle mass all raise creatinine without anything happening to the kidneys.

What is my average blood sugar if my A1C is 6.5%?

About 140 mg/dL (7.8 mmol/L) by the ADAG regression, AG = 28.7 × A1C − 46.7, fitted on 507 subjects with an R² of 0.84. But a point estimate on its own overstates what that equation supports. At an A1C of 7% the paper's own published 95% limits run 123–185 mg/dL around a point estimate of 154, and at 8% they run 147–217 around 183. Two people with the same A1C genuinely can have different average glucose. The regression is also only validated over A1C 4–12%.

Why is there no total cholesterol row on this chart?

Because no source this site opened attributes a normal or optimal total cholesterol to a named population, and the same is true of HDL. The lipid figures that can be sourced are LDL and triglyceride band boundaries from Martin 2013 and treatment thresholds from Sampson 2020, and they are printed as exactly that. It is worth knowing that the US and European classifications differ — the US list cuts at 130 and 160 mg/dL where the European one cuts at 155 — which on its own should discourage anyone from reading a band boundary as a fact about biology. Your LDL is also calculated rather than measured, by one of three equations that can disagree by 12 mg/dL on one blood sample.

Can I use this chart instead of the range on my own lab report?

No, and no printable chart can be used that way. A reference interval belongs to the assay that produced it. Testosterone measured by an uncertified direct immunoassay is not interchangeable with a result calibrated to the CDC reference method, and the harmonized interval on this sheet assumes the latter. An A1C is only comparable if the laboratory uses an NGSP-certified method, which NIDDK states explicitly. Use this sheet to work out what KIND of number you are looking at, who it was measured in and what it assumes about the assay — then read your own figure against your own report, with the clinician who ordered it.

Sources

1. [1]Travison TG, Vesper HW, Orwoll E, et al. Harmonized Reference Ranges for Circulating Testosterone Levels in Men of Four Cohort Studies in the United States and Europe. J Clin Endocrinol Metab. 2017;102(4):1161–1173 (PMID 28324103). Abstract: "In healthy nonobese men, 19 to 39 years, harmonized 2.5th, 5th, 50th, 95th, and 97.5th percentile values were 264, 303, 531, 852, and 916 ng/dL, respectively." and "Harmonized normal range in a healthy nonobese population of European and American men, 19 to 39 years, is 264 to 916 ng/dL."
2. [2]Travison 2017, full text (PMC5460736) — Table 2 (observed percentiles in the 19–39 reference sample) and Table 4 (model-based centiles by decade of age), and the T-score against Z-score discussion behind the decision not to age-adjust the diagnostic range
3. [3]National Institute of Diabetes and Digestive and Kidney Diseases — The A1C Test and Diabetes. Its results table is headed "Diagnosis" and reads "Normal below 5.7 percent / Prediabetes 5.7 to 6.4 percent / Diabetes 6.5 percent or above", footnoted "Any test used to diagnose diabetes requires confirmation with a second measurement, unless there are clear symptoms of diabetes." On the assay: "When using the A1C test for diagnosis, your doctor will send your blood sample taken from a vein to a lab that uses an NGSP-certified method."
4. [4]Nathan DM, Kuenen J, Borg R, et al. Translating the A1C assay into estimated average glucose values. Diabetes Care. 2008;31(8):1473–1478 (PMID 18540046). Abstract: "Linear regression analysis between the A1C and AG values provided the tightest correlations (AG(mg/dl) = 28.7 x A1C - 46.7, R(2) = 0.84, P < 0.0001)" in "A total of 507 subjects, including 268 patients with type 1 diabetes, 159 with type 2 diabetes, and 80 nondiabetic subjects from 10 international centers"
5. [5]Nathan 2008, full text (PMC2742903) — Table 2, the source of the published 95% prediction limits reproduced in the A1c table on this sheet
6. [6]Martin SS, Blaha MJ, Elshazly MB, et al. Comparison of a novel method vs the Friedewald equation for estimating low-density lipoprotein cholesterol levels from the standard lipid profile. JAMA. 2013;310(19):2061–2068 — the source of the US and European LDL-C classification band boundaries and of the two different SI conversion factors for cholesterol and triglycerides
7. [7]Sampson M, Ling C, Sun Q, et al. A New Equation for Calculation of Low-Density Lipoprotein Cholesterol in Patients With Normolipidemia and/or Hypertriglyceridemia. JAMA Cardiol. 2020;5(5):540–548 (PMID 32101259) — the 70, 100 and 190 mg/dL treatment thresholds and the 800 mg/dL triglyceride validation limit
8. [8]Inker LA, Eneanya ND, Coresh J, et al. New Creatinine- and Cystatin C-Based Equations to Estimate GFR without Race. N Engl J Med. 2021;385(19):1737–1749 (PMID 34554658) — the CKD-EPI 2021 creatinine equation behind every eGFR on this sheet, which contains no race coefficient
9. [9]KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney Int. 2024;105(4S):S117–S314 — Table 2, the GFR categories and their terms, with the footnote "In the absence of evidence of kidney damage, neither G1 nor G2 fulfills the criteria for CKD", and Table 14 on the 18-and-over age range of the CKD-EPI equations