Objective
To describe the normative course of maternal sleep during the first 4 months postpartum.
Study Design
Sleep was objectively measured using continuous wrist actigraphy. This was a longitudinal, field-based assessment of nocturnal sleep during postpartum weeks 2 through 16. Fifty mothers participated during postpartum weeks 2 through 13; 24 participated during postpartum weeks 9 through 16.
Results
Maternal nocturnal sleep time was 7.2 (SD ± 0.95) hours and did not change significantly across postpartum weeks 2 through 16. Maternal sleep efficiency did improve across weeks 2 (79.7%; SD ± 5.5) through 16 (90.2%; SD ± 3.5) as a function of decreased sleep fragmentation across weeks 2 (21.7; SD ± 5.2) through 16 (12.8; SD ± 3.3).
Conclusion
Though postpartum mothers’ total sleep time was higher than expected during the initial postpartum months, this sleep was highly fragmented (similar to fragmenting sleep disorders) and inefficient. This profile of disturbed sleep should be considered in intervention designs and family leave policies.
A major barrier to understanding the relations between sleep and postpartum functioning is the lack of normative longitudinal postpartum sleep data on nondepressed postpartum women. A thorough description of normative postpartum sleep is necessary to better understand the relations between sleep and postpartum depression. Specifically, postpartum sleep disturbance has received increasing attention among women’s health researchers, because it has been identified as a precipitating factor for depressive symptoms.
In addition, preventative and intervention efforts to reduce sleep disturbance focus on increasing maternal sleep time, based on a presumption that partial sleep deprivation (an overall attenuation in the total amount of sleep) is to blame for adverse mood effects. This presumption makes conceptual sense, because partial sleep deprivation is known to erode mood, cognitive processes, and psychomotor skills. However, sleep fragmentation is an aspect of sleep disturbance that is generally under addressed within postpartum sleep studies. Furthermore, sleep fragmentation may be equally, or possibly, more important to consider than partial sleep deprivation during the early postpartum period. In contrast to sleep deprivation, sleep fragmentation causes frequent interruptions to sleep architecture throughout the night, although generally preserving total sleep time.
Sleep fragmentation is increasingly studied, because it is a distinct feature of highly prevalent sleep disorders, such as obstructive sleep apnea and periodic limb movement disorder. Consistent with sleep deprivation, sleep fragmentation, whether experimentally induced or the result of a sleep disorder, also leads to excessive daytime sleepiness and decrements in cognitive performance, executive function, and quality of life.
Thus, the cumulative data on partial sleep deprivation and sleep fragmentation support the notion that certain profiles of sleep may not be sufficient to maintain normal functioning. Rather, a minimum duration of uninterrupted sleep may be necessary to maximize sleep’s neurocognitive benefits.
It is reasonable to expect that sleep fragmentation is a primary component of postpartum sleep disturbance. Subjective sleep disturbance is positively associated and sleep effectiveness is negatively associated with maternal fatigue through 6 weeks’ postpartum. New mothers report being surprised by their level of sleep disturbance and daytime exhaustion; mothers describe their sleep as a negotiated behavior in which they strategically adjust their sleep schedules to match their newborns’ polyphasic sleep pattern. Compared with pregnancy, the postpartum period is characterized by a self-report of 3 times the number of nighttime awakenings, a decrease in sleep efficiency, and twice the level of daytime sleepiness. The majority of postpartum mothers’ sleep disturbances are caused by the newborns’ sleep and feeding schedules.
It is appropriate that maternal sleep disturbance should be carefully considered, because it is likely a precursor of postpartum depression, which is a significant health concern. However, it is necessary to first understand normative postpartum sleep profiles, especially sleep efficiency and sleep fragmentation, so that we may better describe the cause of postpartum depression. Our purpose was to provide normative, objectively measured, field-based reference values for nocturnal time in bed, nocturnal sleep time, nocturnal sleep efficiency, nocturnal sleep fragmentation, and diurnal nap frequencies and their durations from weeks 2 through 16 after delivery among nondepressed postpartum mothers with healthy infants.
Quantification of normative postpartum sleep reference values, and specifically, quantification of sleep duration vs fragmentation, were the primary goals of this work. So that our data would have high ecologic validity, we used longitudinal field-based actigraphy among a socioeconomically diverse sample of postpartum women. The data are presented in graphic form as mean and percentiles to broaden their use as reference values. Based on the extant postpartum sleep literature and in the context of known effects of fragmenting sleep disorders, we expected to find that postpartum mothers’ total sleep time would be low, that their sleep would be highly fragmented, and that these effects would remain constant across the first several postpartum months.
Materials and Methods
Participants
The study was approved by the West Virginia University Office of Research Compliance. Women were recruited prenatally via childbirth classes in their third trimester, community advertisements, and word-of-mouth. Telephone screening was conducted before administration of informed consent and health information portability and accountability act authorization. Women were excluded from participation and referred for further evaluation and treatment, as appropriate, on the bases of a history of major depressive or anxiety disorder, a score ≥16 on the Center for Epidemiological Studies of Depression, pregnancy with multiples, premature infant delivery, or infant admission to the neonatal intensive care unit (NICU). All other respondents who were pregnant or whose infant was <1 week old were recruited for participation.
In the first year of the study, we included both primiparous and multiparous women who participated during postpartum weeks 9 through 16 (ie, phase 1). During this time, the project became funded and, based on the success of the initial data collection, we changed the protocol to include only primiparous women who participated during postpartum weeks 2 through 13 (ie, phase 2).
Actigraphy and sleep diary
A member of the research team visited participants’ homes during each week of the study. During each home visit participants were given a new actigraph and personal digital assistant (PDA). Sleep measures were objectively recorded with continuous wrist actigraphy using Mini Mitter’s Actiwatch-64 (AW-64; Mini-Mitter Company, Inc, Bend, OR) actigraphs. An actigraph is a small, nonintrusive wristwatch-like device that collects motion data as the participants follow their normal daily routine. The AW-64 has been validated for recognition of adult sleep based on these movement patterns.
The highest resolution was used (15-second epochs), allowing up to 11 days of continuous recording. Actiware Software version 5.5 (Mini-Mitter Company, Inc) was used to manage, analyze, and archive actigraphy data. The Actiware software uses an algorithm that scores individual epochs as sleep or wake by comparison to a wake threshold value. The validated default wake threshold value parameter setting was used (40).
Periods of nocturnal sleep and daytime naps were participant-identified for analyses using PDA-based sleep diaries (Bruner Consulting, Inc, Longmont, CO) completed in real time at every bed and rise time for nocturnal sleep and diurnal nap periods. The software also has a method for retrospective entries, which are annotated as retrospective. Using the sleep diary, we identified the daytime nap and nocturnal sleep periods reported by the participant. To reduce participants’ burden, we did not ask them to indicate the exact moment of lights out or use the actigraph’s event marker. Thus, we made no attempt to determine sleep-onset-latency and this measure is not possible for analysis. The following measures were analyzed using Actiware software during identified sleep periods:
Time in bed : minutes from the first epoch identified as sleep (the first 2-minute bout of immobility after the diary-indicated bedtime) to the final 2-minute bout of immobility preceding wake closest to diary-indicated risetime.
Total sleep time : minutes of sleep during time in bed .
Sleep efficiency : percent of minutes of sleep during time in bed .
Fragmentation index : sum of percent mobile (percent of epochs during time in bed scored as mobile, including subwake threshold) and the ratio of percent 1-minute immobile bouts to percent mobile (percent of 1-minute bouts during time in bed scored as immobile).
Depression scores
The Edinburgh Postnatal Depression Scale (EPDS) was administered at home visits during postpartum weeks 9 and 12 for phase 1, and during postpartum weeks 2, 4, 6, 8 10, and 12 for phase 2. Each of the 10 items has 4 response options, with a total range = 0–30. Seven of the items are reverse-scored, and the response options vary. A question and answer example is: “I have felt sad or miserable” (Yes, most of the time; Yes, quite often; Not very often; No, not at all). Instrument validation indicated that cutoffs at 9 of 10 and at 11 of 12 have high sensitivity and specificity for identifying possible and probable cases, respectively, of postpartum depression. Participants’ data were entirely excluded from all analyses reported here (but the participant was not dropped from the study) if they had an EPDS score >9 at any administration.
Twenty-two women were excluded from analyses on the basis of an EPDS score >9 during 1 or more weeks during the study. Analyses were based on data from 22 mothers in the first phase (postpartum weeks 9 through 16) and 50 in the second phase (postpartum weeks 2 through 13). In phase 1, one participant withdrew from the study after 4 weeks. In phase 2, five participants withdrew from the study after 4 (n = 2), 5, 7, and 8 weeks, respectively. No participant withdrew from the study because of postpartum depression, and those who were dropped from the current analyses did not have a significantly higher depression scores; therefore, their data recorded before their withdrawal from the study are included in the analyses. In both phases, some data were lost because of actigraph or PDA malfunction, and participant nonadherence to the protocol. Among the 72 postpartum participants (excluding weeks lost because of participant withdrawal from the study), there were 31 weeks of missing data of 745 possible recording weeks (4.2% of possible weeks missing).
Statistical analyses
The size of the first-sample phase was exploratory; the second-phase sample was determined using power analysis for the larger study based on effect sizes from the PI’s previous work with the same actigraphy system. It was determined that (with 2-tailed alpha level = .05 and sigma = 0.80) 43 subjects were required. Our recruitment of 70 participants exceeded that goal, to allow us greater confidence in our ability to generalize for this normative description and protect against attrition.
Data from each participant were averaged within each postpartum week. A minimum of 4 days of recording time were required to calculate each participant’s week average. Data were analyzed with SPSS version 16.0 (SPSS, Inc, Chicago, IL). Descriptive statistics were calculated. One-way analyses of variance (ANOVA) were used to determine statistically significant differences between groups. Linear trend analyses were used to determine statistically significant linear changes across postpartum weeks. The χ 2 analyses were used to evaluate frequency data. Analyses were considered statistically significant when P < .05. Cohen’s d was used to describe effect sizes. Data are shown as mean ± standard deviation (SD).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
