The significance of the phase of circadian rhythmicity for the diagnosis of sleep disturbance was investigated in a group of 80 chronic insomniacs (59 females; mean age 34.8, range 18-59 years). In order to stay close to common clinical practice, data were collected by means of two-week sleep diaries in combination with repeated measurements of subjective alertness and oral temperature. Special measures were taken to minimize the impact of masking upon the temperature measurements. In addition, wrist activity was monitored for an overlapping period of 11 days. Measurements of oral temperature and subjective alertness were fitted with 3(rd)-degree polynomials, for which the peak times (times of maximum) were identified. Principal Components Analysis of these peak times and the times of bed-in and wake-up for all subjects revealed that the phase estimates for the alertness and the sleep-wake rhythms had a strong interrelationship, which was independent from the temperature phase. Using the 25- and the 75-percentiles of the frequency distribution of the temperature peak times as boundaries, the subjects were classified into early (N = 18), middle (N = 37) and late (N = 19) temperature phase subgroups, which had mean peak times of 14:08 h, 17:43 h and 20:09 h, respectively. Comparisons between the early phase and the late phase subgroups showed that a significant overall MANOVA effect was mainly due to differences in total sleep time (early late) as calculated from the log, and to differences in the mean nocturnal actigraphic count (early late). Moreover, the subjective estimates of sleep latency (early late) and wake after sleep onset (early late) tended to differ between the two subgroups. The main result of this study, i.e., that insomniacs with a relatively advanced temperature phase had a relatively shorter and more restless sleep, while insomniacs with a relatively delayed temperature phase tended to experience a relatively long sleep latency, supports the conclusion that the addition of oral temperature measurements to a sleep/wake log extends its diagnostic and therapeutic applicability.
