Swings in glucose tracked with peripheral artery disease across eleven cohorts

Design

Systematic review and meta-analysis of eleven longitudinal cohorts

Population

Adults in cohorts reporting long-term glucose variability and incident PAD

Primary outcome

Risk ratio of peripheral artery disease, high versus low variability

Effect

RR 1.42 (95% CI 1.21-1.66), p<0.001; I-squared 91%

A systematic review and meta-analysis pooled eleven longitudinal cohorts to ask whether long-term glycaemic variability — the swing between readings over months and years, rather than the average — predicts incident peripheral artery disease.

Higher variability was associated with a risk ratio of 1.42 for PAD (95% CI 1.21 to 1.66, p<0.001). The association held across subgroups defined by region, study design, diabetes status, which variability metric was used, how PAD was diagnosed, and whether the analysis adjusted for HbA1c — that last point being the one that matters most, because it suggests variability is carrying information the average does not.

The caution is heterogeneity, and it is substantial: I-squared was 91%. One subgroup difference was significant — cohorts followed for under eight years showed a stronger association than those followed longer (RR 1.64 versus 1.19, p=0.006), which is the pattern you would expect if some of the early signal reflects people who were already unwell.

• Eleven cohorts; high versus low long-term glycaemic variability.
• The association persisted after adjustment for HbA1c, so this is not simply average glycaemia in another form.
• Shorter follow-up gave a stronger signal (RR 1.64 under 8 years versus 1.19 at or beyond), which argues for some reverse causation.

Why it matters

Peripheral artery disease in diabetes is usually found late, and this points at a risk marker already sitting in records you hold.

Don't overread it

This is observational and I-squared was 91%. It establishes an association, not that smoothing variability prevents peripheral artery disease.

The statistics, in plain English

A risk ratio of 1.42 means the high-variability group developed PAD about 42% more often. The confidence interval, 1.21 to 1.66, excludes 1, so the direction is reasonably firm; the width, and the very high heterogeneity, mean the size of the effect is not.

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.