Application of non-contact hematocrit prediction technologies to overcome hematocrit effects on immunosuppressant quantification from dried blood spots

Fully automated dried blood spot (DBS) analysis for therapeutic drug monitoring (TDM) of the immunosup-pressants tacrolimus, sirolimus, everolimus and cyclosporin A suffers from a so-called hematocrit (hct) effect. This effect is related to the analysis of a partial DBS punch and extractability differences imposed by blood with different hcts. As this is intrinsic to automated DBS analysis, this poses a serious drawback for accurate immu-nosuppressant quantification. Knowledge of a sample's hct allows to correct the derived immunosuppressant concentrations for this effect. Unfortunately, when using the DBS approach for sampling at patients' homes, this hct will typically not be available. The aim of this study was to investigate the validity of a correction algorithm during fully automated DBS analysis of immunosuppressants, based on knowledge of the DBS ' hct, obtained via two distinct non-contact hematocrit prediction strategies, using either near-infrared (NIR) or ultra-violet/visible (UV/VIS) spectroscopy.For tacrolimus, sirolimus, everolimus, and cyclosporin A, 48, 47, 58 and 48 paired venous whole blood and venous DBS patient samples were collected, respectively, and analyzed using an automated DBS-MS 500 HCT extraction unit coupled to a liquid chromatography tandem mass spectrometry system. Additionally, for all 201 samples the hct of the DBS was predicted based on NIR and UV/VIS spectroscopy. For tacrolimus and cyclosporin A, both hct prediction strategies allowed for adequate correction of the hct effect. Also for sirolimus and ever-olimus the results greatly improved after hct correction, although a hct bias remained for sirolimus and for everolimus a slightly significant hct effect was observed after NIR-and UV/VIS-based correction. Application of both hct prediction strategies ensured that clinical acceptance limits (i.e. >= 80% of the samples within 20% difference compared to whole blood) were met for all analytes.In conclusion, we demonstrated that non-contact hct prediction strategies, applied in tandem with fully automated DBS analysis, can be used to adequately correct immunosuppressant concentrations, yielding a good agreement with whole blood.