Systematic review and meta-analysis compares non-invasive and interstitial continuous glucose monitoring accuracy

This systematic review and meta-alllysis synthesized data from 38 cohorts, including 20 NIGM and 18 iCGM studies, involving a total sample size of 1,693. The primary objective was to compare the Mean Absolute Relative Difference (MARD) between non-invasive glucose monitoring and interstitial continuous glucose monitoring.

The pooled MARD for NIGM was 10.21% (95% CI: 8.73-11.69%) and for iCGM was 11.82% (95% CI: 10.36-13.29%). There was no significant difference in MARD between the two technologies (p = 0.13). However, the authors noted extreme heterogeneity across studies (I^2 = 95.2%) and identified methodological asymmetries between NIGM and iCGM studies.

Meta-regression identified study duration as the strongest predictor of NIGM accuracy (p < 0.001), with MARD degrading from 8.7% in short-term studies to 15.2% in long-term studies. Additionally, NIGM cohorts were significantly less likely to be validated in the hypoglycemia range compared to iCGM cohorts (15% of NIGM vs 89% of iCGM; p < 0.001).

Clinicians should note that reported NIGM accuracy is substantially influenced by methodological asymmetries and study duration. The high level of heterogeneity and the observed degradation in accuracy over longer durations suggest that current evidence for NIGM accuracy requires cautious interpretation.