Glycaemic Variability

Glycaemic variability is the swing in glucose around its own average, a separate quantity from that average. It is summarised by the standard deviation, by the coefficient of variation, which divides that by the mean and is comparable across people, and by excursion measures such as the mean amplitude of glycaemic excursions. Continuous monitoring added exposure measures, of which time in range, the proportion of readings between 3.9 and 10.0 mmol per litre, is now the most used.

International consensus targets published in 2019 gave the working numbers: time in range above seventy percent, time below 3.9 mmol per litre under four percent, time below 3.0 under one percent, and a coefficient of variation at or below thirty-six percent separating stable from unstable profiles. Mean sensor glucose converts into a glucose management indicator, named to stop people calling it an estimated HbA1c, since the two disagree in individuals because of red cell turnover.

The reason to look at variability separately is that two patients with the same HbA1c can have entirely different profiles, one flat and one alternating between lows and high excursions, and only the second has a hypoglycaemia problem. That said, evidence that reducing variability itself improves hard outcomes is thinner than for lowering mean glycaemia, so it is best treated as a safety descriptor rather than an independent target.

Consumer monitoring is where the term goes wrong. In people without diabetes, post-meal rises into the seven to eight range are normal physiology, not spikes, and sensors with a mean absolute relative difference near ten percent cannot resolve the small between-food differences that personalised nutrition products rank. Compression artefacts during sleep also generate false lows that inflate any variability metric taken from raw traces.