DEXA Z-Score vs T-Score: What's the Difference?

What's the difference between a T-score and a Z-score?

A T-score compares your bone density to the peak density of a healthy young adult; a Z-score compares it to the average for a person of your own age, sex and ethnicity. They appear side by side on most reports, but they answer two different questions. The T-score asks "how do your bones compare to bone at its strongest?" The Z-score asks "how do your bones compare to other people like you, right now?" NHS, 2024

Both scores are measured in "standard deviations" — how far above or below the relevant average your bones sit — so a negative number simply means below that average. The crucial point is that they use different reference groups, which is why the same person can have a worrying T-score and a perfectly normal Z-score at the same time. Which one actually matters for you depends almost entirely on your age and, for women, your menopausal status.

If you'd like the full walkthrough of how a report is laid out first, our companion guide on how to read your DEXA results and T-score covers the headline number in detail. This article zooms in on the second number — the Z-score — and when it takes over.

What is a T-score, and who is it for?

Your T-score compares your bone density to the average peak density of a healthy young adult — and it's the diagnostic number for postmenopausal women and men aged 50 and over. The World Health Organization's reference point is a white female aged 20–29 from the NHANES III dataset, though in practice you're compared to a young adult of your own sex. ISCD, 2023 StatPearls describes the same reference as the mean density of healthy young adults, roughly a 30-year-old. StatPearls, 2023

The T-score falls into one of three WHO bands, as summarised by the Royal Osteoporosis Society and StatPearls: Royal Osteoporosis Society, 2024

• +1 to −1 — normal.
• −1.0 to −2.5 — osteopenia (low bone mass).
• −2.5 or below — osteoporosis.

The International Society for Clinical Densitometry is explicit that T-scores are the preferred measure for postmenopausal women and for men aged 50 and over. ISCD, 2023 For everyone else, it is the wrong tool — and that's where the Z-score comes in.

What is a Z-score, and who is it for?

Your Z-score compares your bone density to the average for a person of your own age, sex and ethnicity — and it's the preferred measure for premenopausal women, men under 50 and children. The NHS describes this comparison as being against "an adult of your own age, gender and ethnicity", in contrast to the young-adult basis of the T-score. NHS, 2024 (For the Z-score specifically, the ISCD notes that self-reported ethnicity is used in the calculation.) ISCD, 2023

The reason the reference group is switched matters. Bone density naturally rises to a peak in early adulthood and then declines with age, so measuring a young person against a young-adult peak is clinically appropriate — but measuring a young person's bones and then, separately, checking them against their own age group tells you whether anything unusual is going on beyond normal development. The ISCD's official position is unambiguous: in children, in premenopausal women, and in men younger than 50, "Z-scores, not T-scores, are preferred". ISCD, 2023 This has been the consensus for two decades — Khan and colleagues wrote in 2004 that "in men under 50 yr of age, premenopausal women, and children, Z-scores, not T-scores, should be used when reporting bone density results". Khan et al., 2004

The Royal Osteoporosis Society puts it plainly: "The Z-score compares your bone density to people of the same age as you," and it is "normally only useful for looking at the bone strength of children and young adults". Royal Osteoporosis Society, 2024

What does a Z-score of −2.0 or lower mean?

A Z-score of −2.0 or lower is defined by the ISCD as "below the expected range for age"; a Z-score above −2.0 is "within the expected range for age". ISCD, 2023 In a younger person, that phrase is doing careful work: it deliberately avoids saying "osteoporosis". A low Z-score is a signal, not a diagnosis.

What it signals is that something other than ordinary ageing may be thinning the bone. StatPearls is direct about the next step: a Z-score "less than −2.0 should prompt a comprehensive evaluation for secondary causes of bone loss". StatPearls, 2023 These secondary causes are common in younger patients — one review of premenopausal women found that between 50% and 90% of those with low bone density had an identifiable underlying cause. Pre-menopausal osteoporosis review, 2010

Common examples of secondary causes include:

• Long-term steroid (glucocorticoid) use — one of the most frequent culprits.
• Eating disorders such as anorexia, and the low body weight and hormonal disruption that come with them.
• Oestrogen deficiency from a range of causes.
• Coeliac disease and other gut conditions that impair nutrient absorption.
• Certain medications, such as some anti-epileptic drugs.

This list is illustrative rather than exhaustive, and the specific cause is something for a doctor to investigate. Pre-menopausal osteoporosis review, 2010

Can your Z-score be normal but your bones still be weak?

Yes — and in older adults this is the Z-score's biggest limitation. Because everyone's bones thin with age, the Z-score compares you against peers who are also losing bone. That means the number can read as reassuringly "normal for your age" while your bones are genuinely weak. The Royal Osteoporosis Society spells out this exact trap: your Z-score "might be normal for your age but your bones may still be weak enough to need a medicine". Royal Osteoporosis Society, 2024

Worked examples: which number matters for whom?

The simplest way to see this is with two people who have identical-looking "low" numbers but completely different next steps. (These are illustrative examples for comprehension, not real cases.)

Imagine a 68-year-old woman with a T-score of −2.6. She is postmenopausal, so the T-score is the number that counts — and −2.6 meets the WHO threshold for osteoporosis (−2.5 or below). Her result is a diagnosis, and the conversation moves to fracture-risk assessment and treatment. Royal Osteoporosis Society, 2024

Now imagine a 34-year-old woman with a Z-score of −2.3. She is premenopausal, so the Z-score is the number that counts — and −2.3 is "below the expected range for age". ISCD, 2023 That does not hand her an osteoporosis label. Instead it triggers a search for a secondary cause — steroids, an eating disorder, oestrogen deficiency, a gut condition — because in her age group a low reading usually reflects something treatable rather than age. StatPearls, 2023 In practice, the usual criteria for actually diagnosing osteoporosis in a premenopausal woman are a Z-score of −2.0 or lower combined with an identified secondary cause or a low-trauma (fragility) fracture — not the density number on its own. Pre-menopausal osteoporosis review, 2010

Same-looking numbers; entirely different meaning — because the reference group is different. If you want a broader tour of everything a scan reports, see our guide to what a DEXA scan shows.

How does this apply to a Screen My Bones scan?

Screen My Bones uses REMS — a radiation-free ultrasound scan — rather than DEXA, but the scoring logic is the same: results are still expressed as a T-score and a Z-score against reference populations, so the interpretation rules in this article carry across. That means the same age-based question applies: if you're postmenopausal or a man over 50, your T-score is the headline; if you're younger or premenopausal, your Z-score is the one to weigh, and a low reading is a reason to investigate rather than to panic.

Because REMS involves no radiation, it's well suited to establishing a baseline you can safely re-measure over time — and the trend between scans often tells you more than any single number. If you'd like a proactive starting point, our bone health assessment is built around exactly that, or you can book a scan directly.

---

This article is part of our screening technology series.