Can your HbA1c also be too low?
An HbA1c below 5.0% is associated with an increased risk of death in people without diabetes, although this is likely a signal of an underlying condition. Always have a strikingly low value investigated by your doctor.
HbA1c is a measure of how high your blood sugar has been on average over the past two to three months. Most attention goes to values that are too high, but a study among people without diabetes shows that a very low value (below 5.0%) is also associated with a higher risk of death. Compared with a value in the middle range (5.0-5.7%), that risk was 30% higher in the short term and 12% higher over ten years. This is a statistical association, not a proven causal relationship. A low value may therefore reflect an underlying condition rather than cause harm itself.
One indication of this comes from brain injury research. An analysis of tens of thousands of lab results showed that brain injury was the only condition in which HbA1c was consistently lower than in healthy individuals. In all 35 other disease categories examined, HbA1c was either elevated or within normal range. A low HbA1c can therefore sometimes be a signal that something else is going on in the body.
In addition, a low measurement can simply be unreliable. Certain conditions or medications, such as an autoimmune disease combined with specific drugs, can interfere with the measurement so that the result appears much lower than the actual blood sugar level. In such a case, a different test, such as fructosamine, is a better option. This comes from a case report, so it is a rare situation, but it does show that a strikingly low result always deserves closer examination.
Finally, there is an observational study in pregnant women suggesting that a low-normal HbA1c reduces the blood sugar buffer. In mothers with an HbA1c below 5.6%, severe gum disease was associated with a one-and-a-half to two times higher chance of having a baby with a low birth weight. In mothers with a slightly higher value, that association disappeared. What explains this mechanism is not yet clear; the study is purely observational and was conducted in a specific Brazilian study population.
All claims are based on associational research (no RCTs). The mortality figures come from one larger study (PMID 33378417). The brain injury research is a large laboratory analysis (PMID 30905458). The falsely low measurement and the pregnancy finding are respectively a case report and an observational study (PMID 34611711, 34792204).