Competing models for the analysis of red cell survival obtained with 51Cr labelling technique

A computer program for the analysis of 51Cr labelled red cell survival data was used in 84 patients and 8 normal subjects to estimate the parameters of 7 competing models derived from different hypotheses of the red cell destruction process. The purpose was to establish the optimal complexity of a red cell survival model to be used. Senescence or random destruction were the most common identified patterns of red cell death (79% of the cases). Models assuming different levels of death probability according to the age of cells, or 2 populations of cells with different death probabilities, were selected in 21% of the cases. These latter patients were characterized by the lowest values of MRCL. The conclusion of the work was that an automatic procedure for selecting the best model of red cell destruction is necessary when information concerning the mechanism of red cell death is required. For the more practical aim of obtaining an accurate measurement of the MRCL, the minimal set of competing models should account of at least 2 different populations of cells in addition to the random or senescence destruction models.