Measurement of 1,5-anhydroglucitol in blood and saliva: from non-targeted metabolomics to biochemical assay

Anna Halama, Michal Kulinski, Sara Abdul Kader, Noothan J. Satheesh, Abdul Badi Abou-Samra, Karsten Suhre, Ramzi M. Mohammad

Diabetes testing using saliva, rather than blood and urine, could facilitate diabetes screening in public spaces. We previously identified 1,5-anhydro-D-glucitol (1,5-AG) in saliva as a diabetes biomarker. The Glycomark™ assay kit is FDA approved for 1,5-AG measurement in blood. Here we evaluated its applicability for 1,5-AG quantification in saliva. Using pooled saliva samples, we validated Glycomark™ assay use with a RX Daytona(+) clinical chemistry analyser. We then used this set-up to analyse 82 paired blood and saliva samples from a diabetes case-control study, for which broad mass spectrometry-based characterization of the blood and saliva metabolome was also available. Osmolality was measured to account for potential variability in saliva samples. The technical variability of the read-outs for the pooled saliva samples (CV = 2.05 %) was comparable to that obtained with manufacturer-provided blood surrogate quality controls (CV = 1.38-1.8 %). We found a high correlation between Glycomark assay and mass spectrometry measurements of serum 1,5-AG (r(2) = 0.902), showing reproducibility of the non-targeted metabolomics results. The significant correlation between the osmolality measurements performed at two independent platforms with the time interval of 2 years (r(2) = 0.887), also indicates the sample integrity. The assay read-out for saliva was not correlated with the mass spectrometry-based 1,5-AG saliva measurements. Comparison with the full saliva metabolome revealed a high correlation of the saliva assay read-outs with galactose. Glycomark™ assay read-outs for saliva were stable and replicable. However, the signal was dominated by galactose, which is biochemically similar to 1,5-AG and absent in blood. Adapting the 1,5-AG kit for saliva analysis will require enzymatic depletion of galactose. This should be feasible, since the assay already includes a similar step for glucose depletion from blood samples.