Is "mild anemia" really mild? Hemoglobin below the normal cutoff linked with triple 10 year mortality.

Here’s a question we rarely stop to ask: when a hemoglobin comes back a little low and we call it “mild anemia,” is it really mild? The number we use to define anemia comes from a World Health Organization definition set back in 1959, based on a relatively small group of healthy young adults. This study asks whether that line still means what we think it means.

Researchers used the UK Biobank, a large population study, and looked at nearly 478,000 adults who had a hemoglobin measured when they enrolled. The median age was 58, and about 54% were women. They followed these people for a median of about 13.6 years and tracked who died – from any cause, and specifically from cardiovascular disease and cancer. Rather than simply labeling people anemic or not, they sorted them by how far their hemoglobin sat above or below the current WHO threshold, which is 13.0 g/dL for men and 12.0 g/dL for women.

The relationship between hemoglobin and death turned out to be U-shaped. The lowest mortality was not at the threshold – it was 1 to 3 g/dL above it. In other words, the healthiest range sits comfortably above the line we currently draw for “normal.” People whose hemoglobin was clearly low, more than 2 g/dL below the threshold, had roughly triple the mortality: their 10-year rate of death was 12.5%, compared with 4.5% in the best-off group, for a hazard ratio of 2.87.

Here’s the part that speaks directly to the “mild” question. Even people whose hemoglobin was technically normal – just 0 to 1 g/dL above the threshold, a value most of us would barely note – still had a measurably higher risk of dying, with a hazard ratio of 1.18. So the risk doesn’t politely wait until the number crosses the official cutoff. It starts creeping up while the value still looks acceptable on the report.

The U-shape cuts both ways. Hemoglobin that was quite high, more than 4 g/dL above the threshold, was also associated with higher mortality – a reminder that more is not better. The associations were strongest and most consistent in men and in older adults, and were less consistent in women under 60.

The bottom line: a mildly low, or even borderline-normal, hemoglobin may not be as harmless as the label “mild anemia” suggests. At the very least, it’s a number that deserves a second look rather than a shrug.

A few important caveats. This is an observational study, so it cannot tell us that low hemoglobin causes earlier death. It’s far more likely that hemoglobin is acting as a barometer – a low value often reflects something else going on, whether that’s iron deficiency from occult blood loss, kidney disease, chronic inflammation, or an undiagnosed cancer. The hemoglobin was measured only once, the population was predominantly White, and, importantly, the authors are explicit that these findings should not be used to redefine the threshold for anemia. This is not a call to start transfusing or treating everyone with a borderline number to some new target.

What I take from it is simpler. The next time a hemoglobin comes back mildly low, or sitting just above the line, it may be worth asking why, rather than filing it under “mild.” That value may be pointing at something we haven’t looked for yet.

Reference: Ahmad, et al. Association between blood hemoglobin concentration and mortality by age and sex: a prospective cohort study. Ann Intern Med 2026 Sep 8; Online ahead of print.