On stationary and non-stationary red cell survival curves.

Abstract This paper presents an extension of a previously reported analysis (Bergner, 1962); at the same time the work is confined to a more specific type of cell-survival data: essentially to the kind of data that can be obtained from studies with DF 32 P-labelled red cells. The analysis is partly motivated by clinical requirements; e.g. it has become desirable to extend the former analyses to the clinically more realistic non-stationary system. The first part of the paper is devoted to conceptual problems in the significance of death probability. An analysis of the stationary system then follows, and a formal model for the normal system is formulated. The non-stationary (non-normal) system is studied in terms of perturbations of the stationary, normal system. Only simple perturbations are studied, but some general conclusions for the further work seem possible. The experimentally most useful result of the analysis is probably a theorem, according to which the initial slope of a survival curve always determines the mean death probability of the cells; the classical concept of mean life span comes out to be less useful under non-stationary conditions than is the mean death probability.