Determination of glucose in plasma by dry film-based near infrared spectroscopy: Correcting the thickness variations of dry films without applying an internal standard

• A model is proposed for the NIR transmission measurements of dry films.
• The DCSSVR calibration strategy was derived from the proposed model.
• It can realize accurate determination of glucose in dried plasmas samples.
• It has great potential in quantitative NIR spectroscopic analysis of dried samples.

The quantitative determination of analytes in plasma by dry film-based near infrared spectroscopy (NIR) is significantly affected by the thickness variations of dry films. Internal standards are generally used to enhance the reproducibility of quantitative results. However, it is difficult to select an appropriate internal standard with general applicability, and the introduction of an internal standard into plasma samples also complicates experimental procedures. In this work, a quantitative NIR transmission spectroscopy model was proposed to explicitly model the effects of the variations in thickness and light scattering characteristics of dry plasma films on NIR transmission measurements, and an advance dual calibration strategy based on support vector regression (DCSSVR) was derived from the proposed quantitative model to realize accurate determination of analytes (e.g. glucose) in dried plasma samples without the use of internal standards. By using the dry film method coupled with the proposed technique, glucose in plasma could be determined over a concentration range of 0.4–20 mmol L− 1 with satisfactory accuracy (average relative predictive error less than 7.0%) while avoiding any use of internal standards. It can reasonably be expected that the DCSSVR strategy might be of major benefit for the quantitative NIR spectroscopic analysis of analytes of clinical significance in dried biofluid samples.