Helios

Methodology

What Helios reasons from, and how much of it is settled

Helios compares your labs against reference intervals. Those intervals are the whole premise of the tool, so they are published here in full — including the ones that no clinician has signed off on. This page is generated directly from the engine's range table, so it cannot quietly fall out of date with what the software actually does.

Review coverage

90
Markers in the engine

Every biomarker Helios can evaluate, numeric and qualitative.

28
Clinician-reviewed

A licensed clinician has read and signed off on the interval.

62
Not yet clinician-reviewed

No clinician has signed these off. Most have a published source behind them; some are genuinely contested.

62 of the 90 reference ranges in Helios have not been signed off by a clinician. Of those, 47 have a citable published source but no clinical sign-off, and 15 are flagged as unsettled — contested, age-dependent, derived from other values, or with no source on record at all. We publish this because a tool that implied all 90 were settled would be lying to you. Findings that rest on an unsettled range are marked as such inside the product, not hidden.

What the three review states mean

Clinician-reviewed

A licensed clinician has read this interval and signed off on it.

Evidence-established

A published source backs this interval, but no clinician has signed it off. Not the same as reviewed.

Needs review

The interval is contested, age-dependent, derived, or has no source on record. Treat it as unsettled.

These are not shades of the same thing. "Evidence-established" means a paper or guideline supports the interval — it does not mean a doctor looked at it in the context of this tool. We keep the two separate on purpose.

Two bands per marker: standard vs optimal

Every numeric marker carries two intervals, and the difference between them matters.

Standard band
The conventional laboratory "normal" — the range that covers most of the general population. Falling outside it is what a typical lab report flags. It describes what is common, which is not the same as what is healthy: it is derived from a population that includes plenty of unwell people.
Optimal band
The tighter target used in longevity and preventive practice — where the evidence suggests risk is lowest, rather than merely where most people sit. This is the band Helios flags against, which is why it will surface things your lab report called normal.

The honest caveat: optimal bands are inherently less settled than standard ones. Many rest on observational association rather than on trials showing that moving a number produces a better outcome. Each marker below carries a direction telling you which side is the concern — higher, lower, or both extremes.

Evidence tiers on recommendations

Anything Helios suggests carries a tier describing how strong the evidence behind it actually is. Tiers are never upgraded to make a suggestion look better supported than it is.

Tier A — trials and guidelines
Backed by randomised controlled trials or established clinical guidelines. The strongest thing we can say.
Tier B — cohort and mechanistic
Backed by observational cohort studies or a plausible biological mechanism. Suggestive, not proven — these show association, and association is not causation.
Tier C — emerging and anecdotal
Early research, small studies, or practitioner experience. Included for completeness and explicitly labelled. Treat with real scepticism.

What this engine will not do

  • It does not diagnose. An out-of-range marker is an observation to discuss with a clinician, never a condition.
  • It does not prescribe. Every medication, hormone, peptide, and therapy it surfaces carries dosing ranges, an evidence tier, and a requires-licensed-clinician flag. It is information to bring to an appointment, not an instruction to follow.
  • The safety gate runs before ranking, always. Interventions blocked on safety grounds — interactions, contraindications, your stated conditions — never enter the recommendation list at all. They are set aside as withheld for safety, and you can see that they were withheld.
  • It does not invent your data. Values parsed out of an uploaded lab PDF are proposals until you confirm them. Nothing extracted by OCR or a language model is trusted automatically, and uncertain values are surfaced as uncertain rather than guessed.
  • It does not send your health data to third parties. Queries to external research databases carry only topic strings such as "testosterone replacement outcomes" — never your identity, your labs, or your symptoms.

Every reference range in the engine

All 90 markers, grouped by body system. Filter by review state to see exactly what is signed and what is not. Markers flagged needs review show the reason their interval is still contested.

Cardiovascular (9)

MarkerStandard bandOptimal bandDirectionReview state
ApoB
mg/dL
<90
<70
Higher is worseClinician-reviewed
Cholesterol/HDL ratio
ratio
<5
<3.5
Higher is worseEvidence-established

Risk rises >5 (Framingham-derived ratio).

HDL-C
mg/dL
Male40–90
Female50–90
Male50–90
Female60–90
Lower is worseClinician-reviewed
LDL-C
mg/dL
<100
<70
Higher is worseClinician-reviewed
Lp(a)
nmol/L
<125
<75
Higher is worseClinician-reviewed
Non-HDL cholesterol
mg/dL
<130
<100
Higher is worseEvidence-established

Non-HDL goals — ACC/AHA / NLA.

Total cholesterol
mg/dL
125–200
150–180
Both extremes are worseEvidence-established

Desirable <200 mg/dL — NCEP ATP-III.

Triglycerides
mg/dL
<150
<80
Higher is worseClinician-reviewed
VLDL
mg/dL
<30
<20
Higher is worseEvidence-established

Calculated VLDL ~2–30 mg/dL (derived from triglycerides).

Hematologic (20)

MarkerStandard bandOptimal bandDirectionReview state
Basophils %
%
<2
<1
Higher is worseEvidence-established

WBC differential reference intervals (reference labs).

Basophils (abs)
x10^3/uL
<0.2
<0.1
Higher is worseEvidence-established

Absolute basophil count reference interval (reference labs).

Eosinophils %
%
<5
<3
Higher is worseEvidence-established

WBC differential reference intervals (reference labs).

Eosinophils (abs)
x10^3/uL
<0.5
<0.3
Higher is worseEvidence-established

Absolute eosinophil count reference interval (reference labs).

Hematocrit
%
Male38.3–50
Female35.5–44.9
Male40–48
Female37–44
Both extremes are worseClinician-reviewed
Hemoglobin
g/dL
Male13.5–17.5
Female12–15.5
Male14–16.5
Female12.5–14.5
Both extremes are worseEvidence-established

WHO anemia cutoffs M<13 / F<12 g/dL; reference labs.

Lymphocytes %
%
20–45
25–40
Both extremes are worseEvidence-established

WBC differential reference intervals (reference labs).

Lymphocytes (abs)
x10^3/uL
1–3.5
1.4–3
Both extremes are worseEvidence-established

Absolute lymphocyte count reference interval (reference labs).

MCH
pg
27–33
28–32
Both extremes are worseEvidence-established

Reference-lab adult interval ~27–33 pg.

MCHC
g/dL
32–36
33–35
Both extremes are worseEvidence-established

Reference-lab adult interval ~32–36 g/dL.

MCV
fL
80–100
85–92
Both extremes are worseEvidence-established

Reference-lab adult interval ~80–100 fL.

Monocytes %
%
2–10
3–8
Both extremes are worseEvidence-established

WBC differential reference intervals (reference labs).

Monocytes (abs)
x10^3/uL
0.1–0.9
0.2–0.7
Both extremes are worseEvidence-established

Absolute monocyte count reference interval (reference labs).

Neutrophils %
%
40–75
45–65
Both extremes are worseEvidence-established

WBC differential reference intervals (reference labs).

Neutrophils (abs)
x10^3/uL
1.8–7.5
2–6
Both extremes are worseEvidence-established

Absolute neutrophil count reference interval (reference labs).

PSA
ng/mL
<6.5age-stratified
<3
Higher is worseClinician-reviewed
Platelets
x10^3/uL
150–400
200–350
Both extremes are worseEvidence-established

Reference-lab adult interval ~150–400 x10^3/uL.

RBC
x10^6/uL
Male4.5–5.9
Female4–5.2
Male4.7–5.5
Female4.2–4.9
Both extremes are worseEvidence-established

Reference-lab adult intervals (StatPearls NBK604207).

RDW
%
<14.5
<13
Higher is worseEvidence-established

~11.5–15%; high RDW associates with mortality (PMC5640961).

WBC
x10^3/uL
3.5–10.5
4–8
Both extremes are worseEvidence-established

Reference-lab adult interval ~4.5–11.0 x10^3/uL.

Hormonal (6)

MarkerStandard bandOptimal bandDirectionReview state
DHEA-S
ug/dL
Male40–290age-stratified
Female15–160age-stratified
Male120–250
Female70–150
Lower is worseClinician-reviewed
Estradiol
pg/mL
Male10–40
Female30–400
Male20–30
Female50–200
Both extremes are worseClinician-reviewed
Free testosterone
pg/mL
Male50–200
Female1–8.5
Male120–200
Female3–7
Lower is worseClinician-reviewed
IGF-1
ng/mL
65–220age-stratified
100–180
Both extremes are worseClinician-reviewed
SHBG
nmol/L
20–60
25–45
Both extremes are worseClinician-reviewed
Total testosterone
ng/dL
Male300–1000
Female15–70
Male600–900
Female30–60
Lower is worseClinician-reviewed

Inflammation (9)

MarkerStandard bandOptimal bandDirectionReview state
ALT
U/L
Male<44
Female<32
Male<25
Female<20
Higher is worseEvidence-established

Healthy ULN ~29–33 (M) / 19–25 (F) U/L — ACG 2017 (PubMed 27995906).

AST
U/L
Male<40
Female<32
Male<25
Female<22
Higher is worseEvidence-established

Lab ULN ~35–40 U/L (Mayo Proceedings LFT review).

Alkaline phosphatase
U/L
40–129
50–95
Both extremes are worseEvidence-established

~30–120 U/L — Cleveland Clinic LFT.

Bilirubin (total)
mg/dL
0.2–1.2
0.3–1
Both extremes are worseNeeds review

'Lower is better' optimal is likely inverted — higher bilirubin associates with antioxidant/protective effects. See docs/RANGE_EVIDENCE.md (fix #6).

ESR
mm/hr
Male<20
Female<30
Male<10
Female<15
Higher is worseNeeds review

Age-dependent (Westergren; Miller upper bound ≈ age/2). Our fixed band is wide and not age-adjusted — no authoritative endorsement. See docs/RANGE_EVIDENCE.md §C.

GGT
U/L
Male<65
Female<45
Male<20
Female<16
Higher is worseClinician-reviewed
Homocysteine
umol/L
<15
<8
Higher is worseClinician-reviewed
Rheumatoid factor
IU/mL
<14
<10
Higher is worseEvidence-established

Normal <14(–20) IU/mL — Cleveland Clinic.

hs-CRP
mg/L
<3
<1
Higher is worseClinician-reviewed

Metabolic (15)

MarkerStandard bandOptimal bandDirectionReview state
Albumin
g/dL
3.5–5
4.3–5
Both extremes are worseEvidence-established

3.5–5.0 g/dL — Medscape lab values.

Albumin/Globulin ratio
ratio
1–2.5
1.2–2.2
Both extremes are worseNeeds review

Derived albumin/globulin ratio; no authoritative optimal interval. See docs/RANGE_EVIDENCE.md.

Bicarbonate (CO2)
mmol/L
22–30
24–28
Both extremes are worseEvidence-established

22–29 mmol/L — Mayo BMP reference.

Calcium
mg/dL
8.6–10.3
9.2–10
Both extremes are worseEvidence-established

8.4–10.6 mg/dL — Medscape lab values.

Chloride
mmol/L
98–107
100–106
Both extremes are worseEvidence-established

98–107 mmol/L — Mayo reference.

Fasting glucose
mg/dL
70–99
75–90
Higher is worseClinician-reviewed
Fasting insulin
uIU/mL
2–19
2–6
Higher is worseClinician-reviewed
Globulin
g/dL
2–3.9
2.4–3.2
Both extremes are worseNeeds review

Calculated (total protein − albumin); upper bound wide with no authoritative optimal interval. See docs/RANGE_EVIDENCE.md.

HOMA-IR
index
<2.5
<1.5
Higher is worseClinician-reviewed
HbA1c
%
4–5.6
4.5–5.3
Both extremes are worseClinician-reviewed
Phosphate
mg/dL
2.5–4.5
3–4
Both extremes are worseEvidence-established

2.5–4.5 mg/dL — Medscape lab values.

Potassium
mmol/L
3.5–5.1
4–4.8
Both extremes are worseEvidence-established

3.6–5.2 mmol/L — Mayo reference.

Sodium
mmol/L
135–145
137–142
Both extremes are worseEvidence-established

135–145 mmol/L — Mayo BMP reference.

Total protein
g/dL
6–8.3
6.5–7.8
Both extremes are worseEvidence-established

6.3–8.3 g/dL — Medscape lab values.

Uric acid
mg/dL
Male3.4–7
Female2.4–6
Male3.5–5.5
Female3–5
Both extremes are worseEvidence-established

Reference intervals; treat-to-target <6.0 mg/dL — ACR 2020 gout (PubMed 32391934).

Micronutrient (8)

MarkerStandard bandOptimal bandDirectionReview state
Ferritin
ng/mL
Male30–400
Female15–200
Male50–150
Female40–120
Both extremes are worseClinician-reviewed
Folate
ng/mL
3–20
10–20
Lower is worseEvidence-established

Deficiency <3–4 ng/mL (NHANES-derived).

Iron
ug/dL
Male65–175
Female50–170
Male90–150
Female70–140
Both extremes are worseEvidence-established

~50–175 µg/dL adult reference (Medscape).

RBC Magnesium
mg/dL
4–6.4
5–6.4
Lower is worseClinician-reviewed
TIBC
ug/dL
250–450
250–400
Both extremes are worseEvidence-established

~250–370(–450) µg/dL (Medscape).

Transferrin saturation
%
20–50
25–40
Both extremes are worseEvidence-established

20–50% (<20 deficient, >50 overload) — Medscape.

Vitamin B12
pg/mL
200–900
500–900
Lower is worseClinician-reviewed
Vitamin D (25-OH)
ng/mL
30–100
50–80
Both extremes are worseClinician-reviewed

Oncologic-Screening (4)

MarkerStandard bandOptimal bandDirectionReview state
AFP
ng/mL
<8
<5
Higher is worseEvidence-established

Adult reference <8 ng/mL (lab-variable).

CA 125
U/mL
<35
<30
Higher is worseEvidence-established

Standard cutoff <35 U/mL.

CA 19-9
U/mL
<37
<30
Higher is worseEvidence-established

Standard manufacturer cutoff <37 U/mL.

CEA
ng/mL
<3
<2.5
Higher is worseEvidence-established

Nonsmoker <3.0 ng/mL — Cleveland Clinic.

Renal (15)

MarkerStandard bandOptimal bandDirectionReview state
BUN
mg/dL
7–20
10–18
Both extremes are worseEvidence-established

6–21 mg/dL — Medscape lab values.

Creatinine
mg/dL
Male0.7–1.3
Female0.6–1.1
Male0.8–1.1
Female0.6–1
Both extremes are worseEvidence-established

M 0.74–1.35 / F 0.59–1.04 mg/dL — Medscape lab values.

Urine RBC
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

Urine WBC
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

Urine bilirubin
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

Urine blood
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

Urine glucose
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

Urine ketones
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

Urine leukocyte esterase
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

Urine nitrite
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

Urine pH
pH
4.5–8
6–7
Both extremes are worseEvidence-established

4.5–8.0 — AAFP urinalysis.

Urine protein
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

Urine specific gravity
SG
1.005–1.03
1.01–1.025
Both extremes are worseEvidence-established

1.005–1.030 — Cleveland Clinic.

Urine urobilinogen
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

eGFR
mL/min/1.73m2
45–120age-stratified
60–120
Lower is worseClinician-reviewed

Serology (1)

MarkerStandard bandOptimal bandDirectionReview state
HIV 1/2 antibody
qualitative
Qualitative (present / absent) — no numeric bandBoth extremes are worseNeeds review

No source on record for this interval.

Thyroid (3)

MarkerStandard bandOptimal bandDirectionReview state
Free T3
pg/mL
2.3–4.2
3.2–4.2
Lower is worseClinician-reviewed
Free T4
ng/dL
0.8–1.8
1.1–1.6
Both extremes are worseClinician-reviewed
TSH
mIU/L
0.4–4.5
0.8–2
Both extremes are worseClinician-reviewed

Generated from packages/engine/helios/reference_ranges.py. The published figures are recomputed from the engine's range table rather than maintained by hand, and a test in the engine suite fails if this page falls behind the code.