Quantitative Methods for the Analysis of Circadian and Episodic Hormone Fluctuations

Abstract

The development of sensitive radioimmunoassays and of high frequency blood sampling techniques has made possible to measure accurately the variations in plasma levels of pituitary and other hormones over the 24-h cycle. Data generated challenged the concept of hormonal homeostasis and introduced the notion that time of sampling is an important factor in defining normal and pathological levels. The investigators of hormonal rhythms are faced with the interpretation of data of a new type, namely, time series where at least two types of temporal variation are present: (1) a circadian periodicity (low frequency variation); and (2) short-term, episodic fluctuations (high frequency variation). Because of the complexity of the data and of their dependence upon time, classical statistical methods currently used in clinical investigation cannot be generally applied. The majority of reports on the 24-h profile of blood constituents have therefore been largely descriptive. The determination of possible correlations between hormonal variations and sleep stages has also been often based on visual examination of the data. The sophistication of the sampling procedures and of the biochemical determinations in plasma is thus far greater than that of the quantitative interpretation of the observations.