Applications
PPG for Diabetes Detection and Glucose Monitoring
PPG cannot directly measure blood glucose concentration, but research shows that diabetes-related vascular changes (arterial stiffness, endothelial dysfunction, autonomic neuropathy) create detectable signatures in PPG waveform morphology and HRV.
How it works
Schematic summary of this guide. Waveforms are illustrative, not measured data.
Diabetes mellitus produces progressive vascular and autonomic changes that manifest in PPG. Type 2 diabetes is associated with increased arterial stiffness (elevated PWV and augmentation index), reduced perfusion index (peripheral microangiopathy), blunted HRV (cardiac autonomic neuropathy), and altered waveform morphology. The second derivative PPG (APG) shows characteristic changes in the b/a ratio and the aging index (AI = (b-c-d-e)/a) in diabetic patients that correlate with HbA1c levels.
However, these models detect the presence of diabetes (a chronic vascular state) and cannot monitor glycemic fluctuations on the timescale relevant for insulin management.
Non-invasive glucose monitoring from PPG remains a highly active research area. Near-infrared spectroscopy (NIRS) at glucose-sensitive wavelengths (1000–1800 nm, outside standard consumer PPG range) has shown in-vitro promise but in-vivo performance is confounded by interstitial fluid delays, skin heterogeneity, and temperature effects. Samsung, Apple, and multiple startups have announced cGM programs using PPG variants, but no device has achieved FDA clearance for non-invasive glucose monitoring as of 2025. For what diabetes studies show when wearable PPG is used around metabolic health rather than to read glucose directly, see this review.
Frequently Asked Questions
Can Apple Watch monitor blood glucose?
No. No consumer smartwatch has FDA clearance for non-invasive blood glucose monitoring. Apple has active research programs and acquired patents in this area, but technical challenges remain significant. Current Apple Watch health features do not include glucose estimation.
How does diabetic neuropathy affect PPG signals?
Cardiac autonomic neuropathy in diabetes reduces HRV across all frequency bands (particularly HF power representing vagal tone) and blunts the normal circadian HR variability pattern. Peripheral autonomic neuropathy reduces the sympathetic vasoconstrictor response, increasing baseline perfusion index but reducing reactivity.
What PPG features are most predictive of diabetes?
The most discriminative features in published models include stiffness index (SI), second derivative wave ratios (b/a, d/a), nocturnal LF/HF HRV ratio, diurnal perfusion index variation, and pulse rate recovery rate after postural change. Ensemble models using 15+ features consistently outperform single-feature approaches.