Longevity calculator
Phenotypic age calculator (Levine's equation)
Enter nine values from a standard blood panel and your age to get Levine's phenotypic age. Enter the unit each analyte was reported in — C-reactive protein in particular, where the equation wants mg/dL and most labs print mg/L. Free, instant, no sign-up.
It must be a fasting glucose. The model was fitted on fasting samples, and a post-meal reading is a different quantity wearing the same name.
Check your lab report before choosing. Every hs-CRP assay and most standard panels report mg/L, which is a tenfold larger number than the mg/dL this equation was fitted in.
Check these against your lab report. With this equation the conversion is where the answer is won or lost: a CRP of 2.0 mg/L read as 0.2 mg/dL is correct, and a CRP of 2.0 mg/L entered as 2.0 mg/dL is wrong by more than two years without anything looking odd.
| Serum albumin | 47 g/L |
| Serum creatinine | 88.4 µmol/L |
| Fasting glucose | 5 mmol/L |
| C-reactive protein | 0.1 mg/dL |
| Lymphocytes (% of white cells) | 30 % |
| Mean cell volume (fL) | 90 fL |
| Red cell distribution width (%) | 13 % |
| Alkaline phosphatase (U/L) | 70 U/L |
| White blood cell count (thousands per µL) | 6 ×10³/µL |
A blood panel is a snapshot your doctor should read. What the coach can help with is the training and food side of the things that move some of these markers.
What phenotypic age is
Levine's phenotypic age combines nine blood markers and your chronological age into a single figure in years. The equation is a Gompertz proportional-hazards model: the ten terms make a linear predictor, that becomes a ten-year mortality score for the reference cohort, and the score is inverted through the age-only mortality curve to land back in years. Weights and units come from Levine et al., Aging (Albany NY) 2018;10(4):573-591, Table 1. The transform comes from the published correction to Liu et al., PLoS Medicine 2019;16(2):e1002760 — the original printing omitted a step and could not be evaluated. What the result says is that this panel resembles the panel of an older or younger person in NHANES III. That is the whole claim.
CRP goes in as mg/dL, not mg/L
This is the single most important thing on the page. Levine's coefficient was fitted on CRP in mg/dL, which is what NHANES III reported, while most modern labs and every high-sensitivity assay report mg/L — a tenfold larger number. The term enters as a natural logarithm, so putting mg/L where mg/dL is expected shifts the linear predictor by 0.0954 × ln(10) = 0.22, which at typical values moves the answer by well over two years. It does not fail loudly; it returns a plausible wrong age. Several public implementations of this equation have it wrong. That is why the unit is a required input here rather than a guess from the magnitude, and why the converted panel is printed back to you.
The mortality score is computed and not shown
The equation cannot reach an age without passing through a ten-year mortality score, so this calculator computes one. It deliberately does not return it. The only sentence anyone would write around that number is "your chance of dying in the next ten years is N percent", and that is not what the number is: it is a property of a cohort of NHANES III participants whose panels looked like yours, over one follow-up window, in one country, decades ago. Attaching it to one reader converts a cohort average into a personal prediction it does not support. The age is the same information with the false precision stripped off.
What the nine markers can and cannot see
Albumin, creatinine, fasting glucose, CRP, lymphocyte percentage, mean cell volume, red cell distribution width, alkaline phosphatase and white cell count. Nothing else. Blood pressure, VO₂ max, smoking, body composition, family history and every other thing that matters to health are outside the equation entirely. And a single blood draw moves: acute illness raises CRP and white cells sharply, hydration shifts albumin and creatinine, and several of these markers vary with the time of day. One panel taken during a chest infection is not a slow-moving property of you.
Where this stops
The model was fitted on NHANES III adults, so below 20 it would be applied to a population it never saw and the calculator refuses. The plausibility bounds on each analyte are deliberately wide — wide enough for dialysis-range creatinine and sepsis-range CRP — because their job is to catch a value entered in the wrong unit or the wrong field, not to decide what counts as a normal result. At the extremes of the range the mortality score saturates and the inverse transform stops being defined; the calculator returns nothing rather than a clamped age. This is not a diagnosis, not a life expectancy, and not a substitute for the clinician who ordered the panel.