Prediction of sodium content in commercial processed meat products using near infrared spectroscopy

The present study evaluated the ability of near infrared transmittance (NIT) spectroscopy (FoodScan, 850-1050nm) to predict sodium (Na) content in commercial processed meat products (n=310) as intact and ground samples. Prediction models were built with all samples spectra and with spectra divided in 5 categories according to the manufacturing meat process. Sodium content (%) was determined using inductively coupled plasma optical emission spectrometry. Modified partial least squares regression for the overall samples showed satisfactory predictive ability for intact (coefficient of determination in cross-validation, R²_CV=0.93) and ground samples (R²_CV=0.95). Despite the low number of samples, good specific prediction models were developed for each commercial meat category. In conclusion, NIT is really promising for at-line application to predict Na in processed meat products which could help industry to accomplish the new labelling regulation.