Why a normal testosterone result can miss the point

What a reference range is

A laboratory reference range is usually the middle 95 percent of a reference population. By definition, 2.5 percent of healthy people fall below it and 2.5 percent above it. Who is in that reference population therefore decides what counts as normal. If the group is heavier, older or less healthy than the previous generation, "normal" moves with it.

For testosterone, that matters. Many labs have built their ranges on whoever happened to be tested, including men with overweight, diabetes or other illness that lower testosterone. A lower limit taken from that group is lower than one taken from healthy young men.

The cleanest range we have

Bhasin and colleagues (JCEM, 2011) took a narrow group: healthy, non-obese men aged 19 to 40 in the Framingham Heart Study, measured with LC-MS/MS, the most accurate method available. The mean was 723.8 ng/dL and the range 348.3 to 1,197 ng/dL. The lower limit of about 12.1 nmol/L comes from healthy young men, not from everyone who happened to be tested.

Travison and colleagues (JCEM, 2017) went a step further and harmonised testosterone measurements across four large cohorts in the United States and Europe, calibrated to a CDC reference standard. For healthy non-obese men aged 19 to 39, the harmonised range was 264 to 916 ng/dL (about 9.2 to 31.8 nmol/L). The two studies differ in the details, but they share the point: a range should be anchored in healthy men and in a standardised method.

Why two labs disagree

The method. Immunoassays and mass spectrometry do not give the same number, especially at low concentrations, where many immunoassays read too high or too low. Two labs can report different values for the same blood, and each one prints its own range.

The time of day. Testosterone peaks in the morning and can be around 20 percent lower in the afternoon in younger men. Guidelines, including the Endocrine Society guideline (Bhasin et al., JCEM 2018), ask for a fasting morning sample, repeated on a second morning before a diagnosis.

SHBG. The binding protein rises with age, so total testosterone can look acceptable while the free fraction has fallen. Harman and colleagues showed this in the Baltimore Longitudinal Study of Aging (JCEM, 2001): the free testosterone index fell faster with age than total testosterone did.

Life right now. Acute illness, poor sleep, a hard training block, weight loss on very little food, alcohol and some medicines such as opioids and glucocorticoids all push the number down. A single result taken in such a period says more about that period than about you.

Generations change the baseline

Danish population surveys analysed by Andersson and colleagues (JCEM, 2007) and Finnish data from Perheentupa (2013) found that men born later had lower testosterone than men of the same age born earlier. That is a cohort effect, not ageing. It means that a range built on today's men may describe a population that is already lower than the one before it.

Age is not the whole explanation

A common reply to a low-normal result is "it is your age". Age does lower testosterone slowly, but the baseline analysis of the European Male Ageing Study (EMAS, Wu et al., JCEM 2008) found that obesity and poor general health were linked to lower testosterone at least as strongly as age. Many healthy older men keep values well within the young range. A low value at any age is a reason to look for a cause, not a reason to stop looking.

What "normal" can and cannot tell you

A number printed as normal describes the men around you today, measured with that lab's method. It is not a verdict on your symptoms. In the same way, a number just below the line is not automatically a disease. The useful question is whether the number, the free fraction and the way you feel point in the same direction.

A sensible next step

Take one fasting morning sample before 10 a.m. with:

1. Total testosterone
2. SHBG and albumin, so free testosterone can be calculated
3. LH and FSH

Repeat a low or borderline result on another morning, ideally at the same lab and with the same method. Write down your symptoms before you get the answer, so the result does not shape how you describe them. Then go through the numbers with a doctor who looks at them together with your history.