Albumin is the most abundant protein in blood plasma and one of the cheapest markers on a standard comprehensive metabolic panel. Most labs report a reference range of 3.5-5.0 g/dL. The longevity literature reads it differently: anything below 4.0 starts to look meaningful, and the band that correlates with the lowest mortality sits at 4.3-5.0. Use the calculator at /tools/biological-age/ to see how your reading feeds the Levine 2018 PhenoAge model.
What is serum albumin?
Albumin is synthesized exclusively by the liver at roughly 12-15 g per day in healthy adults. Half-life in circulation is ~21 days. It carries fatty acids, calcium, hormones, and many drugs; it accounts for ~75% of plasma oncotic pressure. The serum concentration is therefore a composite signal of:
- Hepatic synthesis capacity (liver function).
- Loss into urine (nephrotic-range proteinuria) or gut (protein-losing enteropathy).
- Catabolism rate, which rises in systemic inflammation.
- Plasma volume (dilution lowers the apparent concentration; dehydration raises it).
Because the half-life is 21 days, albumin is not an acute marker. It tracks chronic state, not yesterday's flu.
What is a normal albumin range?
The lab reference range (3.5-5.0 g/dL) is built to flag overt pathology. It is wide and forgiving. The mortality literature is narrower.
Cabrerizo et al. 2015 meta-analyzed 24 studies of older adults and found a graded mortality association: each 0.25 g/dL drop below 4.0 raised all-cause mortality risk roughly 24-47% over 5-10 year follow-up Cabrerizo et al. 2015 . Goldwasser & Feldman 1997 reviewed the broader adult literature and reported similar gradients Goldwasser & Feldman 1997 : every 2.5 g/L decrement (about 0.25 g/dL) raised mortality 24-56% across populations.
A reasonable longevity framing:
- 4.5-5.0 g/dL: optimal, typical of healthy adults under 50.
- 4.3-4.5 g/dL: still good; expected drift after age 60.
- 4.0-4.3 g/dL: monitor, look for inflammation or kidney loss.
- 3.8-4.0 g/dL: investigate. Pair with CRP, eGFR, urine protein, liver panel.
- Below 3.8 g/dL: clinical workup warranted.
How it feeds into PhenoAge
Levine et al. 2018 derived a 9-marker biological-age estimator from NHANES III data (n=9,926, ~10-year follow-up) and validated it on NHANES IV Levine et al. 2018 . The formula assigns albumin a negative coefficient: higher serum albumin lowers calculated phenotypic age.
Concretely, an adult with albumin 4.7 g/dL versus an otherwise-identical adult at 3.9 g/dL will show ~3-5 years of difference in PhenoAge output, depending on the other 8 inputs. The coefficient reflects albumin's combined signal for inflammation, nutrition, and hepatic reserve. Run your own numbers at /tools/biological-age/.
PhenoAge is not the only clock. The original Horvath 2013 multi-tissue methylation clock and the Hannum 2013 blood clock measure different upstream signals (CpG methylation), and the GrimAge clock predicts mortality more sharply still. PhenoAge is the only one of the four that runs on a routine blood panel, which is why it anchors our calculator.
What does low albumin mean?
Three mechanisms compete for explanatory weight when albumin drops:
- Inflammation. Acute-phase response down-regulates hepatic albumin synthesis in favor of CRP, fibrinogen, and SAA. A 1.0 g/dL drop in albumin during sepsis can occur in 48-72 hours despite the 21-day half-life because of redistribution and capillary leak. Chronic low-grade inflammation does the same, slower.
- Synthesis failure. Cirrhosis, chronic hepatitis, severe malnutrition, or kwashiorkor-pattern protein deficit. Rare in well-fed adults; common in advanced liver disease.
- Loss. Nephrotic syndrome (>3.5 g/day urinary protein), severe burns, protein-losing enteropathies (Crohn's, celiac, lymphangiectasia). Rule in or out with a urinalysis.
In otherwise healthy older adults, the dominant signal is usually #1: low-grade chronic inflammation. This is why albumin tracks frailty, sarcopenia, and 5-year mortality so reliably in geriatric cohorts.
What drives it
Modifiable factors with non-trivial effects:
- Resolve underlying inflammation. If hs-CRP is elevated, the work is on CRP, not on albumin directly. See the CRP article for the framework. Lowering chronic inflammation is the most reliable way to nudge albumin upward over months.
- Adequate protein intake. 1.2-1.6 g/kg bodyweight per day in adults over 60. Below 1.0 g/kg, the body has less substrate to maintain albumin synthesis. The classic figure is that albumin starts to fall after ~3 weeks of severe protein restriction (<0.5 g/kg). See protein targets for longevity.
- Liver health. Alcohol cessation in heavy drinkers raises albumin within 4-12 weeks. Treat hepatitis C; weight-loss in MASLD/NAFLD restores synthesis capacity.
- Hydration state. Chronic dehydration falsely elevates albumin; over-hydration falsely lowers it. Both effects are small (<0.2 g/dL) and reverse on rehydration.
Do not try to "supplement" albumin. IV human albumin exists but is reserved for resuscitation in cirrhosis and burns; oral protein powders do not directly raise serum albumin in non-deficient adults.
What does high albumin mean?
Levels above 5.0 g/dL are uncommon and almost always reflect dehydration or laboratory artefact. There is no clinical entity of pathologically high albumin in the way there is for low albumin. Recheck with proper hydration before chasing it.
Cross-marker patterns
Albumin moves with creatinine in interesting ways. A reading of albumin 3.7 g/dL plus creatinine 1.4 mg/dL with proteinuria points to nephrotic syndrome. Albumin 3.6 g/dL plus elevated bilirubin and INR points to hepatic synthesis failure. Albumin 3.8 g/dL plus hs-CRP 8 mg/L plus normal liver and kidney function points to chronic inflammation as the driver. The single number is most useful when read alongside the rest of the panel; see also creatinine and CRP.
How to act on yours
Testing cadence:
- Healthy adult under 50: annual CMP is enough.
- Adult over 60 or with chronic disease: every 6-12 months.
- Drift below 4.0 g/dL: repeat in 4-8 weeks before drawing conclusions, since transient illness can dip the reading.
If two consecutive readings sit below 4.0, the workup is straightforward and inexpensive: hs-CRP for inflammation, urine albumin-creatinine ratio for renal loss, ALT/AST/bilirubin/INR for hepatic synthesis, and a clinical eye for sarcopenia or weight loss. Most clinicians will run this without prompting once they see the pattern.
Counter-view
Albumin's status as a "protein nutrition marker" is overstated in older textbooks. In well-fed adults, albumin tracks inflammation more than it tracks protein intake; targeted feeding studies in chronic illness rarely raise albumin without first lowering CRP. The Levine PhenoAge weighting reflects this composite reality rather than any single causal mechanism, which is also why directly "raising albumin" is not a coherent intervention. If your clinician tells you to eat more protein because your albumin is 3.8, the protein advice is reasonable for other reasons; the albumin number itself will not move much until inflammation does.
