Sleep in late pregnancy predicts length of labor and type of delivery

Transcription

1 American Journal of Obstetrics and Gynecology (2004) 191, 2041e6 Sleep in late pregnancy predicts length of labor and type of delivery Kathryn A. Lee, RN, PhD, FAAN,* Caryl L. Gay, PhD Department of Family Health Care Nursing, University of California, San Francisco, San Francisco, Calif Received for publication February 5, 2004; revised May 20, 2004; accepted May 25, 2004 KEY WORDS Sleep Actigraphy Pregnancy Labor Obstetric Cesarean section Objective: The purpose of this study was to test the hypothesis that fatigue and sleep disturbance in late pregnancy are associated with labor duration and delivery type. Study design: In a prospective observational study of 131 women in their ninth month of pregnancy, objective (48-hour wrist actigraphy) and subjective (sleep logs and questionnaires) measures were used to predict labor outcomes using analysis of variance and logistic regression. Results: Controlling for infant birth weight, women who slept less than 6 hours at night had longer labors and were 4.5 times more likely to have cesarean deliveries. Women with severely disrupted sleep had longer labors and were 5.2 times more likely to have cesarean deliveries. Fatigue was unrelated to labor outcomes. Conclusion: Health care providers should prescribe 8 hours of bed time during pregnancy to assure adequate sleep and should include sleep quantity and quality in prenatal assessments as potential predictors of labor duration and delivery type. Ó 2004 Elsevier Inc. All rights reserved. Sleep disturbance and fatigue are common complaints among pregnant women and have been well documented in previous research. 1-6 In other clinical populations, sleep disturbance and fatigue are associated with adverse physical and psychologic outcomes, 7,8 but sleep disturbance and fatigue related to adverse perinatal outcomes have not been adequately studied. Prenatal sleep deprivation has been implicated in the development of postpartum blues, 9 but there is currently no evidence that fatigue or poor sleep is related to labor This work was funded by National Institutes of Health grant no. 1RO1 NR45345 from the National Institute of Nursing Research. * Reprint requests: Kathryn A. Lee, RN, PhD, FAAN, University of California-San Francisco, School of Nursing, Box 0606, San Francisco, CA kathryn.lee@nursing.ucsf.edu outcomes. In a study of 99 women, Evans et al 10 reported no relationship between self-reported sleep quality the week before labor and the duration of labor, type of delivery, or maternal perceptions of labor. However, their conclusions were based on an unvalidated sleep questionnaire, and parity was not controlled. As part of a larger randomized clinical trial to test an intervention to improve sleep for new mothers and fathers during the neonatal period, we obtained both objective and subjective measures of their sleep. Sleep and fatigue data from both the control and intervention groups of primiparae collected during their initial prenatal assessment were used in this analysis related to labor and delivery outcomes. These data were used to test the hypothesis that sleep disturbance and fatigue in late pregnancy are associated with labor duration and type of delivery /$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi: /j.ajog

2 2042 Lee and Gay Material and methods This study was approved by the institution s Committee on Human Research (CHR). As part of a larger randomized clinical trial, expectant couples were recruited from childbirth education classes. Eligible couples included those expecting their first child, both at least 18 years of age, both willing to participate, and both able to read and write English. Couples were excluded if they planned to hire a live-in nanny or if either parent had a diagnosed sleep disorder or worked the night shift. Because of potential sleep disturbance related to the trauma of a prior pregnancy loss, expectant mothers with a history of involuntary pregnancy loss were also excluded. After informed consent was obtained from each individual, couples were assessed in their homes during their last month of pregnancy and during the first 3 months after delivery. All couples were paid for their participation. This article reports on data collected from the women in both the control and experimental groups during their last month of pregnancy before introducing the intervention and subsequent data collected about their birth experience. To objectively estimate sleep quality and quantity, each participant was asked to wear a wrist actigraph (Ambulatory Monitoring, Inc, Ardsley, NY) for 48 hours during her last month of pregnancy. Data were collected only on weekdays to minimize the variability caused by potential differences in weekday and weekend sleep patterns. The wrist actigraph provides continuous motion data by using a battery-operated wristwatch-size microprocessor that senses motion with a piezo-electric linear accelerometer. Actigraph data were analyzed by trained research assistants blinded to pregnancy outcomes with the use of the autoscoring program for sleep available in Action3 software (Ambulatory Monitoring, Inc). The autoscoring algorithm yielded 2 sleep-related outcomes variables: (1) sleep quantity as total sleep time (TST) at night, and (2) sleep quality as wake after sleep onset (WASO). As an estimate of sleep disruption, WASO is reported as the percentage of minutes awake divided by minutes in bed after falling asleep. During a typical 7- to 8- hour sleep period, WASO of 15% represents more than an hour of wake time after falling asleep. WASO between 5% and 10% is typical in healthy, nonpregnant women 5 and greater than 15% was considered severe sleep disruption for this study. Congruence between polysomnographic measures and actigraphy measures indicate adequate validity and reliability when sleep is assessed in healthy young adults, including women of childbearing age with 88% agreement between the 2 methodologies. 14 Previous actigraphy studies have recommended 7-day monitoring periods, but to decrease systematic error caused by changing weekday and weekend sleep patterns and to minimize missing data, a 48-hour weekday monitoring interval was used in this study. While wearing the wrist actigraph, participants used 48-hour sleep logs to record their bed times, wake times, and ratings of their sleep quality. The sleep logs were used to facilitate interpretation of the actigraphy data and also yielded 2 self-report variables of sleep quality and quantity: time in bed (computed from bed times and wake times) and 5-point sleep quality ratings from 1 (very poor) to 5 (very good). Participants completed the General Sleep Disturbance Scale (GSDS) as a subjective measure of their prenatal sleep disturbance. The GSDS contains 21 items rating the frequency of specific sleep problems during the past week from 0 (not at all) to 7 (every day). The GSDS also yields subscale scores addressing sleep quality, sleep latency, sleep quantity, sleep maintenance, early awakening, use of medication to promote sleep, and the impact of sleepiness on daytime function. The scale yields a total score ranging between 0 and 147 and subscale scores between 0 and 7, with higher scores indicating greater frequency of sleep disturbance. Cutoff scores of 43 for the total scale, and a mean of 3 on any one subscale, distinguishes good and poor sleepers. 15 A mean score of 3 or more would indicate that sleep was perceived as disturbed on 3 or more nights during the past week, corresponding to DSM-IV for primary insomnia. 16 The GSDS total score and 3-item sleep quality subscale score were used as variables in this study. In the current sample, the Cronbach alpha coefficient was.82 for the 3-item sleep quality subscale and.80 for the total scale. To estimate perception of fatigue severity, participants also completed a 7-item numerical rating version of the Visual Analog Scale for Fatigue 17 each morning and evening during the 48-hour assessments. In the current sample, the 7-item version had Cronbach alpha coefficients of.91 for evening data and.95 for morning data. After delivery and before the first postpartum sleep assessment at 1 month, women were asked to provide information about their birth experience, including infant s birth weight, duration of labor, and type of delivery. Infant birth weight was reported in pounds and ounces and converted to grams. Labor duration was defined for them as the time from onset of regular contractions to the time of birth. Delivery type was categorized as spontaneous vaginal, assisted vaginal (forceps or vacuum), or cesarean. Measures that were completed twice during the 48- hour assessment (actigraphy, log data of bed times and wake times, ratings of sleep quality, and morning and evening fatigue ratings) were averaged to obtain mean values. Pearson and point-biserial correlations were used to determine the strength of relationship between the independent variables, labor duration, and delivery type. Analysis of covariance (ANCOVA) was used to detect differences in labor duration among sleep cate-

3 Lee and Gay 2043 gories. Logistic regression was used to determine the significance of sleep and fatigue variables in predicting type of delivery. Infant birth weight, maternal age, work status, number of days between sleep assessment and delivery, and group assignment (control or experimental) were evaluated as potential covariates. The data were analyzed with SPSS for Windows version 11.5 (SPSS, Inc, Chicago, Ill). A 2-tailed alpha level of.05 was used for all statistical tests. Results Of the 152 women enrolled in the larger study, 5 women were excluded from analysis because labor onset occurred before completion of the 48-hour sleep monitoring. Six women were excluded because their actigraph monitor malfunctioned, 6 were excluded because they had incomplete self-report data, and 4 were excluded because they had scheduled cesarean births and delivered by cesarean section before the onset of labor. The remaining 131 women had a mean age of 32.1 G 4.5 years. The sample was 67% white, 16% Asian, 9% Hispanic, 6% mixed or other, and 2% black; 82% were college educated; 76% were employed, and 51% were working at the time of the 48-hour prenatal assessment. Given the selection criteria of the larger study, all the women were partnered. Women in the sample were of moderate to high socioeconomic status, with a median household income of $60,000 to $89,999. Prenatal sleep and fatigue data were collected an average of 22.6 G 10.1 days before delivery (range = 3-55 days). In the last month of pregnancy, 21% reported leg twitching or jerking, and 37% reported loud snoring. These sleep disturbances are common in pregnancy, and loud snoring was associated with greater sleep disruption, but neither was related to labor and delivery outcomes. Because of the timing of study enrollment in the last month of pregnancy, all infants were term (36-42 weeks gestational age). Sixty-two percent of the women had spontaneous vaginal births, 17% had assisted vaginal births, and 21% had cesarean births. Sleep variables did not differ between women who had spontaneous vaginal deliveries and those who had assisted vaginal births. Therefore, these women were grouped together for analysis. Two women delivered twins; all others were singleton births. Because the twin births were not outliers on any study variable or in any analysis, they were not excluded. Birth weights ranged from 2330 to 5341 g (mean = 3544 G 513 g). Of the potential covariates, only birth weight was related to both the labor and delivery outcomes and the sleep measures. Birth weight was dichotomized at 4000 g and included as a covariate in the ANCOVA and logistic regression analyses. As expected, there were no differences between the control and experimental groups Table I Correlations between birth outcomes, demographics, and prenatal sleep and fatigue (n = 131) Measure Labor duration* Cesarean birth y Demographic factors Infant birth weight.32 z.21 x Maternal age Actigraphy (2 nights) WASO.18 x.36 z TST ÿ.27 z ÿ.19 x Sleep log (2 nights) Time in bed ÿ Sleep quality ÿ.11 ÿ.06 Self-reported sleep (1 wk) GSDS total GSDS sleep quality x Self-reported fatigue (2 d) Morning fatigue ÿ.08 ÿ.03 Evening fatigue ÿ *Pearson correlations. y point-biserial correlations. z P!.01. x P!.05. on any of the baseline sleep measures or labor and delivery outcomes, and therefore the groups were combined for analysis. Although working women reported spending less time in bed, feeling more fatigued, and having better sleep quality than nonworking women, work status was not associated with labor duration or cesarean delivery and therefore was not included as a covariate in final analyses. Women delivering within 1 week of completing the sleep assessment did not differ on any sleep or fatigue variable from women delivering within 2 weeks or within 3 weeks, and therefore timing of the sleep assessment relative to delivery was not included as a covariate. Maternal age was unrelated to labor and delivery outcomes and was not included as a covariate. Prenatal sleep and labor duration Because self-reported duration of labor was not normally distributed in this sample, the labor durations of the 6 subjects who exceeded 40 hours were truncated to 40 hours, resulting in a more symmetric distribution. The resulting mean labor duration was 19.8 G 12.4 hours. Pearson correlations indicated that labor duration was related to infant birth weight and to actigraphy measures of sleep quality (wake after sleep onset- WASO) and quantity (total sleep time-tst), but unrelated to self-report measures of sleep or fatigue (Table I). Women in this sample had a mean WASO of 13.0% G 7.4% and a mean TST of 7.1 G 1.1 hours. To compare the duration of labor for women with different degrees of sleep disruption and with different

4 2044 Lee and Gay Table II Differences in labor duration by total sleep time and wake after sleep onset (n = 131) Labor duration* n (mean G SD) WASO 15%C G 11.7 y 10%-14.9% G 12.9!10% G 14.6 TST!6 h G 12.5 y h G C h G 15.6 *Adjusted for infant birth weight. y significantly longer labor than the other 2 groups. amounts of sleep, women were grouped into categories based on their percent WASO and their TST (Table II). In healthy, nonpregnant populations, WASO less than 10% is typical, and WASO of 15% or higher is indicative of severe sleep disruption; TST of 7 to 8 hours per night is recommended for most adults, and less than 6 hours per night represents severe sleep restriction. In absence of norms for pregnant women, these parameters were used for the current study. Controlling for infant birth weight, ANCOVA indicated that the degree of prenatal sleep disruption had a significant effect on labor duration (F[2,125] = 6.93, P =.001). Sidak-adjusted post-hoc comparisons indicated that women who had severe sleep disruption (WASO R 15%) had significantly longer labors than women with less disrupted sleep. Women who experienced significant sleep disruption did not compensate for their disrupted sleep by spending more time in bed; time in bed was similar (8.7 G 1.0 hours) for women with and without severely disrupted sleep. Controlling for infant birth weight, ANCOVA indicated that prenatal sleep quantity (TST) also had a significant effect on labor duration (F[2,125] = 5.54, P =.005). Sidak-adjusted post-hoc comparisons indicated that women who averaged less than 6 hours of sleep per night had significantly longer labors than women who averaged 6 hours or more (Table II). Because women who slept less than 6 hours may have been trying, but unable, to get more sleep, WASO and time spent in bed were compared with those who slept 6 hours or more. Although women who averaged less than 6 hours of sleep per night experienced significantly more sleep disruption (19.9% G 9.3%) than women who slept 6 hours or more (11.7% G 6.3%; separate variance t[21.0] =3.77, P =.001), they also spent significantly less time in bed (7.9 G 1.5 hours) than women who slept 6 hours or more (8.8 G 0.8 hours; separate variance t[19.9] = 2.43, P =.025). Time in bed was computed from bed times and wake times recorded in women s sleep logs. An average of 8.7 G 1.0 hours in bed was recorded for each night. Their Table III Differences in cesarean rates by wake after sleep onset (n = 131) n Cesarean rate sleep quality was rated an average of 3.1 G 0.8 (corresponding to a description of fair on the 1 to 5 scale in the sleep log). Neither of these sleep log variables was associated with labor duration. Women in this sample had a mean GSDS total score of 46.3 G 16.1 and a mean sleep quality subscore of 3.7 G 1.8. Participants had a mean morning fatigue rating of 3.8 G 1.7 and a mean evening fatigue rating of 5.8 G 1.5. GSDS scores and fatigue ratings were not associated with labor duration. Prenatal sleep and type of delivery Odds Ratio* 95% CI P value WASO 15%C % y 10%-14.9% % !10% % ref TST!6 h % z h % z 7C h % ref GSDS sleep quality Poor sleep % z 5C d/wk Poor sleep % z 3-4 d/wk Poor sleep 0-2 d/wk % ref *Adjusted for infant birth weight. y P!.01. z P!.05. Like labor duration, delivery type was associated with infant birth weight and actigraphy measures of sleep quality and quantity (Table I). Adjusted odds ratios are presented in Table III. Controlling for infant birth weight, women with severe sleep disruption (WASO R 15%) were 5.2 times more likely to have a cesarean delivery than women who had little or no sleep disruption (WASO! 10%). In addition to disrupted sleep, insufficient total sleep time was also related to delivery type. Again controlling for infant birth weight, women who averaged less than 6 hours of sleep per night were 4.5 times more likely to have a cesarean delivery than women who averaged at least 7 hours of sleep, and women who averaged between 6 and 7 hours of sleep per night were 3.7 times more likely to have a cesarean delivery than women who averaged at least 7 hours of sleep. Unlike labor duration, type of delivery (vaginal vs cesarean) was also associated with women s selfreported sleep quality on the GSDS. Compared with

5 Lee and Gay 2045 women who reported having poor sleep quality 0 to 2 days per week and controlling for infant birth weight, women who reported having poor sleep quality 3 to 4 days per week were 4.2 times more likely to have a cesarean delivery, and women who reported having poor sleep quality 5 or more days per week were 5.3 times more likely to have a cesarean delivery (Table III). Type of delivery was not associated with daily sleep log measures, total GSDS score, or prenatal fatigue ratings in the evening or morning. Comment In this sample of childbearing women, measures of sleep quality and quantity were associated with both labor duration and type of delivery, whereas measures of evening and morning fatigue were not. Women who had less sleep or had severely disrupted sleep, as assessed by actigraphy monitoring, had significantly longer labors and were more likely to have cesarean deliveries than women who slept more and had less disrupted sleep. These findings contradict the conclusions drawn by Evans et al 10 that a woman s prenatal self-assessment of sleep may be relevant to her well-being, but unlikely to influence her labor and delivery. The use of objective measures in the current study yielded consistent relationships between sleep disturbance and labor outcomes. Whereas Evans et al 10 included multiparous women in their sample, this sample was limited to primiparous women. Given the added challenges of caring for other children during pregnancy, and the well-known effects of parity on labor duration, parity should be controlled in sleep research on this population. In this study, infant birth weight was associated with labor outcomes. Women with larger infants had longer labors and were more likely to have cesarean deliveries. Previous research suggests that infant birth weight is associated with dystocia and delivery type. In a study of more than 14,000 births, Poma 18 found that dystocia increased with birth weight, and cesarean rates were higher for infants less than 2500 g or more than 4000 g. The current study had no preterm or low-birth weight infants, but did have 24 infants more than 4000 g. Because this was a study of partnered primiparous women who were somewhat older than the population average primiparous woman, and all delivered healthy term infants, findings cannot be generalized to multiparous or primiparous women who are young, unpartnered, or delivering preterm or small-for-gestational-age infants. Furthermore, most women in this sample had moderate to high socioeconomic status, and therefore the findings cannot be generalized to low-income women whose living conditions may affect their sleep quality, and quantity, as well as labor and delivery outcomes. Finally, the findings of this study are limited by the use of women s reports of labor duration rather than a vaginal examination to distinguish length and stage of labor. The results of this study need to be replicated in other samples of childbearing women. Given the current findings, future studies should include standardized objective measures of sleep quality and quantity and should not rely solely on self-report measures. Previous research has examined the relationship between childbirth outcomes and prenatal factors such as stress hormones 19 and anxiety. 20 Future research should include such physiologic and psychologic factors, because they may help to further explain the relationship between sleep disturbance and labor outcomes. Other factors that may affect labor duration, such as epidural use and induction of labor, should also be included in future research. Given the results of this study, it is recommended that clinicians discuss sleep quantity and quality with their pregnant patients. Although objective measures are ideal, clinicians may obtain adequate estimates of sleep quantity during a clinic visit by asking women about their bed time and wake time the previous day and calculating their time in bed. Following up with questions about how typical the previous night s sleep was, in terms of sleep quality and quantity, and whether her sleep differs on weekdays and weekends could help reduce some of the error associated with generalizing from an assessment of a single night s sleep. For women who seem to be experiencing significant sleep disruption that affects daytime functioning 3 or more days each week, clinicians should encourage ample opportunity for sleep by suggesting an adjustment in bed times and wake times. The women in this study who averaged less than 6 hours of sleep per night also experienced severe sleep disruption. Although severe sleep disruption might explain limited sleep quantity, women who averaged less than 6 hours of sleep per night also spent less time in bed, thereby limiting their opportunity for sleep. Women with severe sleep disruption may need to increase their time in bed to assure a sufficient amount of sleep. Given these findings, health care providers should consider a specific prescription or recommendation of at least 8 hours in bed at nightdfrom 10:00 pm to at least 6:00 am. Similar to the advice that a pregnant woman should be eating for 2, health care providers should likewise stress the importance of sleeping for 2 during prenatal health care visits. Acknowledgments We gratefully acknowledge the technical support from Margaret Taffe, the statistical support from Dr Steven Paul, and the commitment of the research team that included Shih-Yu Lee, Annelise Gardner, Maria Cho,

6 2046 Lee and Gay Claudia Rocha, Naomi Schoenfeld, Valerie Tobin, and Suzanne Towns. References 1. Schweiger MS. Sleep disturbance in pregnancy. Am J Obstet Gynecol 1972;114: Hertz G, Fast A, Feinsilver SH, Albertario CL, Schulman H, Fein AM. Sleep in normal late pregnancy. Sleep 1992;15: Lee KA, DeJoseph JF. Sleep disturbances, vitality, and fatigue among a select group of employed childbearing women. Birth 1992;19: Suzuki S, Dennerstein L, Greenwood KM, Armstrong SM, Satohisa E. Sleeping patterns during pregnancy in Japanese women. J Psychosom Obstet Gynaecol 1994;15: Lee KA, Zaffke ME, McEnany G. Parity and sleep patterns during and after pregnancy. Obstet Gynecol 2000;95: Mindell JA, Jacobson BJ. Sleep disturbances during pregnancy. J Obstet Gynecol Neonatal Nurs 2000;29: Briones B, Adams N, Strauss M, Rosenberg C, Whalen C, Carskadon M, et al. Relationship between sleepiness and general health status. Sleep 1996;19: Spiegel K, Leproult R, Van Cauter E. Impact of sleep debt on metabolic and endocrine function. Lancet 1999;354: Wilkie G, Shapiro CM. Sleep deprivation and the postnatal blues. J Psychosom Res 1992;36: Evans ML, Dick MJ, Clark AS. Sleep during the week before labor: relationships to labor outcomes. Clin Nurs Res 1995;4: Ancoli-Israel S, Cole R, Alessi C, Chambers M, Moorcroft W, Pollak CP. The role of actigraphy in the study of sleep and circadian rhythms. Sleep 2003;26: Jean-Louis G, von Gizycki H, Zizi F, Fookson J, Spielman A, Nunes J. Determination of sleep and wakefulness with the actigraph data analysis software (ADAS). Sleep 1996;19: Walsh JK, Schweitzer PK, Anch AM, Muehlbach MJ, Jenkins NA, Dickins QS. Sleepiness/alertness on a simulated night shift following sleep at home with triazolam. Sleep 1991;14: Cole RJ, Kripke DF, Gruen W, Mullaney DJ, Gillin JC. Automatic sleep/wake identification from wrist activity. Sleep 1992;15: Lee KA. Self-reported sleep disturbances in employed women. Sleep 1992;15: American Psychiatric Association. Diagnostic and statistical manual for mental disorders. 4th ed. (DSM-IV). Washington, DC: APA; p Lee KA, Hicks G, Nino-Murcia G. Validity and reliability of a scale to assess fatigue. Psychiatry Res 1991;36: Poma PA. Correlation of birth weights with cesarean rates. Int J Gynecol Obstet 1999;65: Sandman CA, Wadhwa PD, Chicz-DeMet A, Dunkel-Schetter C, Porto M. Maternal stress, HPA activity, and fetal/infant outcome. Ann N Y Acad Sci 1997;814: Johnson RC, Slade P. Obstetric complications and anxiety during pregnancy: is there a relationship? J Psychosom Obstet Gynaecol 2003;24:1-14.