Reliable analysis of individual human melatonin profiles by complex cosinor analysis

The analysis of diurnal secretion patterns of the pineal hormone melatonin should provide information about magnitude of peak concentration, time of peak (acrophase), and duration of elevated hormone levels. We report here on a method for analysis of human melatonin secretory rhythms which fulfills these requirements with considerable precision. Blood samples were obtained from 10 healthy male volunteers at 30 min intervals throughout a 24 hr period, three times (the initial blood sampling was repeated after 2 weeks and 3 months). These sets of 48 plasma samples per volunteer were analyzed for melatonin by radioimmunoassay. The mean intra-assay variation was 10.3%, and inter-assay variations were 15.8% at 21 pg/ml, 11% at 28 pg/ml, and 9.4% at 48.6 pg/ml. Data were subjected to single cosinor analysis (SCA) or analyzed by a complex cosinor analysis (CCA) with the fundamental and the first harmonic as parameters. Both methods provided essentially the same information about the acrophases, whereas the CCA had a clear advantage in terms of better regression coefficients between the original data points and the calculated curve (CCA: r = 0.952 +/- 0.018; SCA: r = 0.867 +/- 0.039 [means +/- SD]; P < 0.001). As a consequence, maximum and minimum values and the times of onset and cessation of melatonin production could be estimated easily and reliably without the need for rough and/or subjective measures. By comparing the three sampling sessions, the secretory rhythm of each individual was clearly reproducible (mean coefficient of variation 8.4%), thus confirming earlier work. Interindividual differences, however, were quite pronounced, especially with respect to amplitudes (more than 25% coefficient of variation).(ABSTRACT TRUNCATED AT 250 WORDS)