Changes in sleep–wake cycle during the period from late pregnancy to puerperium identified through the wrist actigraph and sleep logs

INTRODUCTION

Since the 1970s, compact types of actigraph that can be placed on the wrist or ankle have been used for evaluating sleep or wake in various fields. It has been reported that there is a high coincidence between the number of activities recorded by actigraph placed on the wrist and the results from sleep–wake evaluation by polysonograph (PSG).^1–4 We also reported that the coincidence between both subjects was 96.3%,⁵ using the algorithm described by Cole et al.³

Actigraphs are used increasingly in clinical aspects such as investigation of depression and apnea syndrome during sleep as well as reduced sleeping quality due to time differential, and are used also for infants and young children. However, there have been only a few studies on the application of the actigraph to pregnant or postpartum women. In this study, we measured the number of actigraphic activities continuously and compared the results with those from sleep logs recorded concomitantly.

METHODS

The subjects were one multiparous woman (26 years old) and three primi gravidae (27, 30, and 31 years old). Their actigraphic activities were measured almost continuously, starting at 34 weeks of gestation until 3 months after delivery. Concomitantly, the subjects were instructed to record the sleep logs. An actigraph (AMA-32CL, Ambulatory Monitoring Inc.) was initialized under the sleep/wake mode with zero crossing mode, 1 min of epoch time and setting the amplifier at 18, and was placed on the non-dominant arm. The subjects were instructed to push an event-marker of the actigraph at the time of sleep or wake. Their sleep and wake conditions obtained from the actigraph were evaluated by means of the algorithm of Cole et al.³ The sleep log was written in terms of various living activities every 10 min regularly at the same time daily.

The data were collected by dividing the pregnancy and postpartum courses into nine sections (in weeks), having the day of delivery as a reference point. Statistical analyses were performed by averaging the daily sleep parameters determined from the actigraphic activities for each week of the pregnancy and postpartum courses of each woman. Also, the results were subjected to one-way analysis of variance and a corresponding t-test was performed by utilizing the data obtained from the 34–37th weeks of pregnancy, during which the sleep parameters were relatively stable, as references.

RESULTS

Figure 1 shows the actigraphic activity pattern during the pregnancy and postpartum courses of one case. During the period from delivery to the fifth postpartum week, the sleep–wake cycle characteristically became greatly irregular and the wake time during the night increased markedly. Table 1 shows the mean and standard deviation of the sleep parameters evaluated by actigraphy in each week of the pregnancy and postpartum courses. The results of the one-way analysis of variance showed there were significant changes in the sleep efficiency and the wake after sleep onset (WASO) during the pregnancy and postpartum periods. When compared to the 34–37th weeks of pregnancy, the sleep efficiency was significantly shorter or shorter during the entire postpartum period. Also, when compared to the 34–37th weeks of pregnancy, WASO was significantly longer during the entire postpartum period. During the second week of postpartum period, the sleep efficiency was lowest, and the WASO was longest, but they both began to recover slightly after that. Most waking hours at night were spent by breast-feeding or taking care of babies. When compared to the 34–37th weeks of pregnancy, the total sleep time was significantly shorter during the first and second weeks of the postpartum period, and the sleep latency was significantly shortened through the first to third postpartum weeks.

Table 1. . Mean and standard deviations of sleep variables from late pregnancy to postpartum for four women

On the other hand, the various sleep parameters obtained from the sleep logs revealed changes similar to those evaluated by the actigraphy. The correlation coefficient between the sleep logs and the sleep parameters obtained by the actigraphy in all subjects was calculated. The results showed significantly high correlation coefficients regarding the time in bed (r = 0.81, P < 0.001); total sleep time (r = 0.71, P < 0.001); sleep efficiency index (r = 0.74, P < 0.001); sleep latency (r = 0.64, P < 0.001); WASO (r = 0.69, P < 0.001); and the total daytime napping time (r = 0.72, P < 0.001). However, relatively low correlation coefficients were seen regarding the number of WASO (r = 0.21, P < 0.01) and the number of naps (r = 0.49, P < 0.001).

DISCUSSION

In the previous study using PSG, the sleep efficiency decreased significantly during the early stage of postpartum period when compared to the late stage of pregnancy.⁶ This tendency was also confirmed by the results of the present study using actigraphy. However, in the present study, WASO increased the most during the second postpartum week, and the sleep efficiency decreased the most during the second postpartum week, but they both began to recover gradually after that. The reason for this phenomenon may be that the lactation cycle to neonates and their sleep–wake cycle began to be synchronized with the day–night cycle. Except for the number of WASO, the results of actigraphy correlated strongly with the contents of the sleep logs.