How to Calculate Free Testosterone With SHBG and Albumin

Free testosterone is calculated from three blood measurements: total testosterone, sex hormone-binding globulin (SHBG), and albumin, plugged into an equation based on how these proteins compete for testosterone in your bloodstream. The most widely validated version of this equation, developed by Vermeulen and colleagues, uses the known binding strengths of SHBG and albumin to estimate how much testosterone is left unbound and biologically active. You don’t need to do the math by hand; several online calculators handle it, though the equation you choose and the accuracy of the inputs both matter more than most people realize.

Why Total Testosterone Falls Short

Only about 2-3% of testosterone in your blood is truly free, meaning it floats unattached to any protein. The rest rides along bound to either SHBG (which grips it tightly) or albumin (which holds it loosely). Total testosterone measures all of it: free, loosely bound, and tightly bound together. That number works fine as a screening tool most of the time. But when SHBG levels are unusually high or low, total testosterone can paint a misleading picture of what your body actually has available to use.

This is a well-recognized problem in endocrinology. Total testosterone does not sufficiently describe androgen status when SHBG is altered.¹ Someone with high SHBG might show a normal total testosterone level while their free testosterone is actually low, because SHBG is locking up more of it than usual. The reverse happens too: a person with low SHBG might have a normal or low-normal total testosterone but plenty of free testosterone floating around. Free testosterone assessment resolves these ambiguities, and the calculation using SHBG and albumin is the most practical way to get there for most clinicians.

What the Calculation Does

The underlying idea is straightforward, even if the math gets involved. Testosterone molecules in your blood are in a constant tug-of-war between SHBG, albumin, and being free. SHBG binds testosterone tightly with a high affinity. Albumin binds it weakly but is present in very large quantities, so it still absorbs a sizable fraction. The calculation uses the law of mass action to model this competition: given the total amount of testosterone, the amount of SHBG, the amount of albumin, and the known binding strength of each protein for testosterone, you can solve for the fraction that ends up unbound.

To run the calculation, you need three lab values:

Total testosterone: typically reported in ng/dL or nmol/L
SHBG: reported in nmol/L
Albumin: reported in g/dL or g/L (some calculators use a default value if you don’t enter one)

The equation also requires two constants: the association constant for testosterone binding to SHBG, and the association constant for testosterone binding to albumin. Different research groups have used slightly different values for these constants, which is one reason different calculators can give you different answers for the same inputs. The output is your estimated free testosterone concentration, sometimes also expressed as a percentage of total testosterone.

Which Equation Performs Best

Several equations have been published over the years, and they do not all agree. A comparison of four different published formulas found that mean biases ranged from about 6% to 56% relative to each other, with some equations consistently overestimating and others underestimating.² That is not a trivial spread when clinical decisions depend on whether someone falls above or below a cutoff.

The gold standard for measuring free testosterone directly is equilibrium dialysis. It is slow, expensive, and not widely available, so calculated free testosterone serves as the practical alternative. The key question is which calculation comes closest to equilibrium dialysis results. A large validation study using samples from the European Male Aging Study compared three main approaches head-to-head against equilibrium dialysis. The Vermeulen equation produced a median ratio of calculated to directly measured free testosterone of 1.19, meaning it ran about 19% high on average. But critically, that ratio stayed stable across different SHBG, testosterone, and albumin levels. The competing equations (by Ly-Handelsman and Zakharov) showed much worse consistency: their ratios were strongly affected by SHBG levels, meaning the calculation drifted substantially depending on the patient’s SHBG. For rank correlation with equilibrium dialysis, the Vermeulen equation scored 0.89, compared to 0.74 for the Ly equation and 0.62 for the Zakharov approach.³

A separate study in women confirmed the broader pattern: calculated free testosterone using the mass action approach correlated almost perfectly with equilibrium dialysis (r = 0.99), while the older analog immunoassay method for measuring free testosterone directly had unacceptable bias and random variability.⁴ This is an important distinction: the cheap “direct” free testosterone assays that some labs offer are not the same thing as a properly calculated free testosterone, and they perform far worse.

Still, even the best equation is imperfect. Regardless of which formula was used, at least a quarter of samples in one study showed what researchers considered unacceptable deviation from equilibrium dialysis, and the worst-performing samples tended to be those with high SHBG concentrations.⁵ The Vermeulen equation is the best practical option available, but it is an estimate, not a direct measurement.

Does Albumin Actually Matter in the Calculation?

Many online calculators either ask for an albumin value or silently assume a default of around 4.3 g/dL. This raises a practical question: do you actually need a measured albumin, or can you just use the default?

For most people, the default works fine. Albumin is present in much higher concentrations than SHBG, but its binding affinity for testosterone is very weak. Because those two factors largely cancel each other out, modest changes in albumin concentration don’t move the needle much on calculated free testosterone. In hyperandrogenic women with albumin levels in the normal physiological range of 4.0 to 5.0 g/dL, researchers found it unnecessary to measure individual albumin concentrations to calculate free and bioavailable testosterone from total testosterone and SHBG.⁶

The exception is when albumin is genuinely low. Accuracy of the calculation decreased when albumin dropped to 3.5 g/dL or below, particularly when SHBG was also low (at or below 30 nmol/L). That combination occurred in only about 1.2% of the study population, but when it did happen, the default assumption introduced meaningful error.⁷ Low albumin can show up in liver disease, kidney disease, severe malnutrition, or critical illness, so if any of those apply, using a measured albumin rather than a default is worth the effort. For everyone else, the standard assumption is reliable enough that the extra lab test adds cost without adding useful precision.

Why SHBG Varies So Much Between People

SHBG is the variable that drives most of the action in the free testosterone calculation, so understanding what moves it around helps you interpret results. SHBG is made predominantly in the liver, and its production is influenced by a surprisingly wide range of factors including obesity, lipid metabolism, inflammation, and genetics.⁸

At the liver level, sex steroids and thyroid hormone are the primary regulators. Estrogen tends to push SHBG up, which is why women on oral contraceptives often have dramatically elevated SHBG. Thyroid hormone does the same, so hyperthyroidism raises SHBG and hypothyroidism lowers it. Meanwhile, circulating pro-inflammatory mediators are strongly correlated with lower SHBG.⁹ This connection to inflammation helps explain why conditions involving chronic low-grade inflammation, like obesity and metabolic syndrome, reliably suppress SHBG.

Fatty liver disease is a particularly potent influence. In a population study of over 1,600 people, men with even mild fatty liver had about double the risk of low SHBG concentrations. Postmenopausal women with fatty liver showed a similar pattern, with about a 2.4-fold increased risk of low SHBG. Beyond liver fat, age, BMI, waist-to-hip ratio, and HDL cholesterol levels all independently influenced SHBG.¹⁰

The practical takeaway is that two people with the same total testosterone can have very different free testosterone levels based on their metabolic health, body composition, thyroid status, and liver function. This is precisely why the free testosterone calculation exists, and why SHBG is the most important input after total testosterone itself.

When Free Testosterone Changes a Clinical Diagnosis

The real-world impact of calculating free testosterone becomes clearest in situations where SHBG is expected to be abnormal. In men, this matters most when evaluating possible testosterone deficiency. When symptoms are present and total testosterone is clearly low, the diagnosis is usually straightforward. But in equivocal cases, measuring or calculating free or bioavailable testosterone can further clarify the picture.¹¹ Free testosterone levels correlate with clinical signs of androgen deficiency better than total testosterone does, particularly in men with sexual symptoms.¹²

The stakes of skipping the calculation can be surprisingly high in certain populations. In a study of HIV-infected men with sexual symptoms, relying on total testosterone alone would have identified only about 11% of patients as hypogonadal. Adding calculated free testosterone and a full hormonal workup revealed that total testosterone alone would have missed roughly half of all diagnoses.¹³ HIV treatment commonly raises SHBG, which masks the deficiency if you only check total testosterone.

In women, the calculation is especially useful for detecting excess androgens. Conditions like polycystic ovary syndrome (PCOS) and idiopathic hirsutism often present with symptoms of androgen excess even when total testosterone falls within the normal reference range. In women with clinically evident hyperandrogenism, calculated free testosterone identified the excess in roughly 89-95% of patients, while total testosterone alone caught only about 36-67%, depending on the diagnostic group.¹⁴ The reason is that women with PCOS often have suppressed SHBG, particularly if they carry extra weight, which lowers the binding protein and amplifies the free fraction even when total levels look unremarkable.

In one study comparing hirsute women and women with PCOS to controls, the directly measured hormone parameters showed no significant differences between groups, while calculated free and bioavailable testosterone were significantly elevated in both clinical groups.¹⁵ Obesity compounds this effect further. In women with hirsutism or PCOS who had a BMI of 25 or above, SHBG was significantly lower and calculated free testosterone was significantly higher compared to their leaner counterparts with the same conditions.¹⁶

Getting the Inputs Right

The accuracy of any free testosterone calculation depends entirely on the accuracy of the numbers you feed into it. A few practical details are worth keeping in mind.

Testosterone itself has a strong diurnal rhythm, peaking in the early morning and declining throughout the day. This is why clinical guidelines generally recommend drawing blood before 10 a.m. for testosterone testing. SHBG, by contrast, shows much less diurnal variation, so the timing concern is driven almost entirely by the testosterone measurement itself.¹⁷ If your blood was drawn in the afternoon, your total testosterone may read lower than your true morning baseline, and the calculated free testosterone will be low as well, potentially giving a false impression of deficiency.

The specific assays your lab uses for total testosterone and SHBG also matter. One study found that while calculated free testosterone correlated almost perfectly with equilibrium dialysis overall, the agreement depended strongly on which assays were used for total testosterone and SHBG.⁴ If you are tracking free testosterone over time, ideally use the same lab and the same calculator each time, so that any changes you see reflect real biological shifts rather than methodological drift.

If your clinician orders a “free testosterone” directly from a standard commercial lab, check whether it was calculated or measured. Many large reference labs use analog immunoassays for “direct” free testosterone measurement, and these perform poorly. Calculated free testosterone using the Vermeulen equation is consistently more accurate and is the preferred approach alongside equilibrium dialysis, which remains the true gold standard but is rarely ordered outside research settings.

The Binding Model Is Simpler Than Reality

Every widely used free testosterone calculator treats SHBG as though it has a single binding site with a fixed affinity for testosterone. The real biology is more complex. SHBG circulates as a dimer with two potential binding sites, and experimental work has demonstrated that the binding process involves multiple steps, including at least two interconverting states of unbound SHBG and allosteric interaction between the two binding sites. A model that accounts for this complexity produced free testosterone estimates that matched equilibrium dialysis in men and reduced systematic error in women compared to the simpler Vermeulen model.¹⁸

Other factors that the standard calculation ignores include temperature, competing steroids, and certain medications. The affinity of SHBG for testosterone decreases as temperature rises, which is an important consideration for laboratory technique but not something you need to worry about as a patient.¹⁹ And certain compounds can competitively displace testosterone from SHBG, potentially increasing the free fraction beyond what the calculation would predict.²⁰ These nuances rarely change clinical decisions for individual patients, but they help explain why no calculated estimate will ever perfectly match a direct measurement for every sample.

Do Genetics Affect the Calculation’s Accuracy?

SHBG has well-known genetic variants that influence circulating SHBG levels. A natural concern is whether these variants also throw off the relationship between calculated and measured free testosterone, perhaps by producing SHBG molecules with altered binding properties that the equation doesn’t account for. A study examining common SHBG gene polymorphisms found that while certain variants did shift SHBG and total testosterone concentrations, they did not significantly alter the agreement between calculated and directly measured free testosterone. Carriers of these variants showed a similar but nonsignificant trend toward slightly different free testosterone levels, but the gap between calculated and measured values was not affected.²¹ In other words, the common genetic variants change how much SHBG you make, but the SHBG itself still binds testosterone in the way the equation expects. The calculation remains valid even if your SHBG levels are genetically higher or lower than average.

Bioavailable Testosterone and the Free Androgen Index

When reading lab reports or online discussions, you will encounter two other terms that are related but distinct from calculated free testosterone. Bioavailable testosterone refers to the free fraction plus the albumin-bound fraction combined. Because albumin holds testosterone loosely enough that tissues can strip it away fairly easily, some researchers argue bioavailable testosterone is a better reflection of what is actually accessible to your body than the strictly free fraction alone. The same Vermeulen-style mass action approach can produce both numbers; the calculator simply adds the albumin-bound portion back in for the bioavailable result.

The free androgen index is a cruder alternative: total testosterone divided by SHBG, multiplied by 100. It requires no albumin value and no complex math. In women, particularly when evaluating possible androgen excess, the free androgen index performs reasonably well as a screening tool and detected hyperandrogenism at rates comparable to calculated free testosterone in several studies.¹⁴ In men, however, the free androgen index is considered unreliable and is not recommended for clinical use. The relationship between the index and actual free testosterone breaks down at male-range testosterone concentrations, making it a poor substitute for the full calculation.