Glucose variability and peripheral artery disease: a real association, badly estimated

Design

systematic review and meta-analysis of eleven longitudinal cohort studies, random-effects model, PROSPERO registered

Population

cohorts of adults with and without diabetes, Asian and Western, with repeated long-term glycaemic measures

Primary outcome

risk ratio of incident peripheral artery disease, high versus low long-term glucose variability

Effect

RR 1.42 (95% CI 1.21–1.66), p<0.001, I² = 91%

Eleven cohorts were pooled to ask whether long-term glucose variability — the swing in HbA1c or fasting glucose across years, not the within-day variability a continuous glucose monitor reports — predicts peripheral artery disease. Participants with high variability had a risk ratio of 1.42 (95% CI 1.21 to 1.66) against those with low variability.

The association survived adjustment for HbA1c in the subgroup analysis, which is the finding that would matter if it held: it would mean variability carries vascular risk beyond average glycaemia. But heterogeneity was I-squared 91%, which is very high, and the effect was substantially larger in cohorts followed for under eight years (RR 1.64) than in those followed longer (RR 1.19, p for difference 0.006).

That pattern is what you see when sicker people at baseline drive early events, not when a risk factor accumulates damage over time. A genuine causal exposure usually shows the opposite gradient. The honest reading is that glucose variability marks something, probably a combination of treatment instability, comorbidity and adherence, and that this analysis cannot say the variability itself is the culprit.

• Check pedal pulses and ask about claudication in anyone whose HbA1c has been swinging between visits.
• Erratic HbA1c is usually a signal about the patient's circumstances — cost, supply, insulin technique, depression — before it is a signal about the vasculature.
• Do not add a variability target to a patient's plan on the strength of this; no trial has shown that smoothing variability changes outcomes.
• Ankle-brachial index is cheap and under-used in Indian diabetes clinics; this is a reason to do more of them, not a new indication.

Why it matters

Variability metrics are increasingly displayed on dashboards and reports; this is a reminder that displaying a number is not the same as having something to do about it.

Don't overread it

These were observational cohorts — they cannot show that reducing variability reduces peripheral artery disease.

The statistics, in plain English

I-squared of 91% means almost all the variation between studies is real disagreement rather than chance, so the pooled 1.42 is an average of results that do not agree with each other — the confidence interval around it looks tighter than the evidence deserves. The subgroup finding cuts the other way from causation: if variability caused peripheral artery disease, longer exposure should produce a bigger effect, and here longer follow-up produced a smaller one.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.