Association of stress-related sleep disturbance with psychiatric symptoms among pregnant women

Abstract

Background:

Physiological changes during pregnancy are often accompanied by reduced sleep quality, sleep disruptions, and insomnia. Studies conducted among men and non-pregnant women have documented psychiatric disorders as common comorbidities of insomnia and other sleep disorders. However, no previous study has examined the association between stress-related sleep disturbances and psychiatric disorders among pregnant women.

Methods:

This cross-sectional study included a total of 2,051 pregnant women in Peru. The Spanish-language version of Ford Insomnia Response to Stress Test (FIRST-S) was used to assess sleep disruptions due to stressful situations. Symptoms of antepartum depression, generalized anxiety disorder, and posttraumatic stress disorder (PTSD) were examined using the Patient Health Questionnaire-9, Generalized Anxiety Disorder Scale-7 and PTSD Checklist – Civilian Version, respectively. High risk for psychosis was assessed using the Prodromal Questionnaire. Multivariable logistic regression procedures were used to estimate adjusted odds ratios (ORs) and 95% confidence intervals (CI).

Results:

Stress-related sleep disturbance was reported by 33.2% of women. Of all women, 24.9% had antepartum depression, 32.2% had generalized anxiety disorder, 30.9% had PTSD, and 27.6% were assessed as having a high risk of psychosis. After adjusting for confounders, women with stress-related sleep disturbances were more likely to experience antepartum depression (OR=2.74; 95%CI: 2.22–3.38), generalized anxiety disorder (OR=2.48; 95%CI: 2.04–3.02), PTSD (OR=2.36; 95%CI: 1.93–2.88), and high risk for psychosis (OR=2.07; 95%CI: 1.69–2.54) as compared to women without stress-related sleep disturbances.

Conclusions:

Stress-related sleep disturbances during pregnancy are associated with increased odds of psychiatric disorders. Inquiring about stress related sleep disturbances during antenatal care may be beneficial for identifying and caring for women at high risk of psychiatric disorders.

Introduction

Physiological, social and behavioral changes during pregnancy are known to contribute to poor sleep quality¹. In a recent meta-analysis, approximately 45.7% of pregnant women were found to experience poor sleep with significant decreases in sleep quality throughout pregnancy.² Furthermore, previous studies have shown antepartum sleep disturbances are associated with adverse pregnancy outcomes including gestational hypertension ^3,4, pre-eclampsia ⁵, abnormal duration of labor ^6,7, mode of delivery ^7–9, and preterm birth ^10,11. The high prevalence of maternal antepartum sleep disturbances and their association with a myriad of important obstetrical and perinatal outcomes warrant further research to examine modifiable factors contributing to disturbances in sleep during pregnancy.

Stress-related sleep disturbances may be clinically useful as screening tools for detecting women at high risk for developing psychiatric disorders during pregnancy. Studies conducted among men and non-pregnant women have shown the comorbidity and predictive value of stress-related sleep disturbances with depression ^12,13, anxiety ¹⁴, and PTSD ^15,16, and psychosis ¹⁷. Little is known, however, about the associations between stress-related sleep disturbances and psychiatric disorders among pregnant women. Furthermore, most previous studies on psychiatric disorders among women are conducted in high income countries – this despite the fact that the burden of maternal depression, anxiety and PTSD is substantially higher in low- and middle-income countries (LAMICs) ^18–21. Given the (1) high burden of psychiatric disorders in LAMICs, and (2) absence of evidence on the association between sleep disturbances and psychiatric disorders during pregnancy, we sought to examine the association of sleep disturbances due to stressful situations with antepartum depression, generalized anxiety disorder, PTSD, and psychosis among a cohort of Peruvian pregnant women.

Methods:

Study population

Participants from the Pregnancy Outcomes, Maternal, and Infant Study (PROMIS) were included in this present study. PrOMIS, a prospective cohort study of pregnant women, was designed to examine maternal, social, and behavioral risk factors in the development of adverse maternal and perinatal outcomes among women attending prenatal clinics at the Instituto Nacional Materno Perinatal (INMP) in Lima, Peru. The INMP is operated by the Peruvian Ministry of Health and is the primary reference establishment for maternal and prenatal care in Peru. The PROMIS cohort has been described previously ^22,23. Participants were eligible if they were at least 18 years old, able to speak and read Spanish, and were less than 16 weeks of gestational age. All participants provided written informed consent. Study procedures were approved by the institutional review boards of the INMP, Lima, Peru and the Office of Human Research Administration at the Harvard T.H. Chan School of Public Health, Boston, MA, USA.

Ford Insomnia Response to Stress Test (FIRST)

The Ford Insomnia Response to Stress Test (FIRST) is a widely used instrument to assess the likelihood of insomnia and sleep disruptions following stressful conditions ^24,25. The FIRST)was used to assess the likelihood of sleep disruptions due to stressful situations and broadly-described periods of stress during the day or evening ²⁴. Participants rated how likely they had difficulty sleeping when they had recently experienced the following 9 items: 1) before an important meeting the next day; 2) after a stressful experience during the day; 3) after a stressful experience in the evening; 4) after getting bad news during the day; 5) after watching a frightening movie or television show; 6) after having a bad day at work; 7) after an argument; 8) before having to speak in public; and 9) before going on vacation the next day. The response categories were scored on a Likert scale from 0 to 4 with the response categories “not likely,” “somewhat likely,” “moderately likely,” and “very likely.” Total scores range from 9 to 36. Two clinically significant FIRST cutoff values were identified in a prior study FIRST ≥ 16 (sensitivity 77%; specificity 50%) and FIRST ≥ 18 (sensitivity 62%; specificity 67%)²⁶. For the purposes of our analysis, we defined presence of stress-related sleep disturbances (i.e., sleep reactivity) as FIRST ≥ 18. We also performed the analysis with the FIRST ≥ 16 cutoff and the results remained virtually unchanged. Furthermore, the Spanish-language version of the FIRST questionnaire has been found to have excellent reliability and good construct validity in this study population²⁷.

Depression

A Spanish-language version of the Patient Health Questionnaire-9 (PHQ-9) was used to evaluate maternal antepartum depressive symptoms ²⁸. The PHQ-9 is a 9-item screening instrument that assesses depression in the 14 days prior to evaluation. PHQ-9 score is calculated by assigning scores of 0–3 to the following response categories: “not at all” “several days” “more than half the days” and “nearly every day.” Depression was defined as a score ≥ 10. The Spanish-language PHQ-9 has been previously validated in Spanish-language and among pregnant populations women ^29–31.

Generalized anxiety disorder

Generalized anxiety disorder was assessed using a Spanish-language version of the Generalized Anxiety Disorder Questionnaire (GAD-7). The 7-item GAD-7 examines the following symptoms in the 14 days prior to screening: 1) nervousness; 2) inability to stop worrying; 3) excessive worry; 4) restlessness; 5) difficulty in relaxing; 6) easily irritated; and 7) fear of something awful happening. GAD-7 score is calculated by assigning scores of 0–3 to the response categories of “not at all” “several days” “more than half the days” and “nearly every day.” GAD-7 total score ranges from 0 to 21. A cutoff of 10 or higher was used to assess presence of anxiety among participants. The GAD-7 has previously been validated among pregnant Peruvian women ³².

Posttraumatic stress disorder

Posttraumatic Stress Disorder (PTSD) was evaluated in our population using a Spanish-language version of the PTSD Checklist – Civilian Version (PCL-C). ³³ The PCL-C includes 17 items that reflect Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria ³⁴. Participants were asked about how bothered they were by each item over the past month on a 5-point Likert scale with regard to their most significant life event stressor. Total scores on the PCL-C range from 17 to 85. We used a PCL-C score of 26 or higher to indicate presence of PTSD in our population. The PCL-C cutoff has previously been validated in this population with a 86% sensitivity and 63% specificity in diagnosing PTSD ³³.

Psychosis

The 16-item Prodromal Questionnaire (PQ-16) assessed the presence of psychotic symptoms ³⁵. The PQ-16 assesses lifetime symptoms of psychosis-like experiences on a two-point scale (true/false). PQ-raw scores are the total number of true responses on the questionnaire. Participants with scores ≥ 6 were classified as high risk for psychosis. The PQ-16 also assesses degree of distress caused by each item on a Likert scale from 0 to 3. Total scores were summed by totaling distress ratings for each experience. An average distress score was calculated by dividing a participant’s PQ-total by PQ-raw score, to provide a measure of the average degree of distress per item endorsed. Average distress had a range of zero to 3. The psychometric properties of the Spanish-language version of the PQ-16 has previously been examined in this population and found to have a good construct validity ³⁶.

Additional covariates

Trained interviewers used structured questionnaires to collect information from participants on sociodemographic and pregnancy characteristics. Participants’ age was categorized as follows: 18–19, 20–29, 30–34, and ≥ 35 years. Other sociodemographic variables included were race/ethnicity (Mestizo vs. others), marital status (married or living with partner vs. others), years of education (≤6, 7–12, and >12 years), employment status (yes vs. no), difficulty accessing basic foods (hard vs. not very hard) and difficulty accessing medical care (hard vs. not very hard). Pregnancy-related characteristics included were parity (nulliparous vs. multiparous), planned pregnancy (yes vs. no), tobacco smoking before or during pregnancy (yes vs. no), alcohol consumption during pregnancy (yes vs. no), and gestational age at interview (in weeks). Early pregnancy body mass index (BMI) (kg/m2) was calculated from the participants’ weight measured to the nearest 0.1kg and height measured with a telescopic height instrument to the nearest 0.1cm and categorized as: <18.5, 18.5–24.9, 25–29.9, ≥30 kg/m².

Analytic population

A total of 2,068 participants were eligible for inclusion in the study. We excluded 6 participants due to missing information on the FIRST questionnaire, 6 on PHQ-9, 2 on GAD7, 5 on PCL-C, and 5 on PQ-16. A total of 2,051 participants remained in the present analysis.

Statistical analysis

We first examined sociodemographic and pregnancy characteristics using number (n) and percentages (%) for categorical variables and mean and standard deviation (SD) for continuous variables. For categorical values, p-values were calculated using the Chi-square test. For continuous variables, p-values were calculated using analysis of variance (ANOVAs). Multivariable logistic regressions were used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for the associations between stress-related sleep disturbances and maternal psychiatric symptoms. Potential a priori confounders (i.e., maternal age, education, and parity) were included in the models. All statistical analyses were performed using IBM SPSS Statistics 25.0 (SPSS Inc, Chicago, Illinois).

Results

Sociodemographic, pregnancy, and psychiatric characteristics of the population are presented in Table 1. The mean age of women was 27.92 years (standard deviation (SD)=6.28), and mean gestational age at interview was 11.10 (SD=3.95) weeks of gestation. The majority of women were Mestizos (a race/ethnicity of mixed Amerindian and European descent; 80.6%), had more than 12 years of education (56.1%), were married or living with a partner (82.7%), and were not employed (50.6%). Difficulty in accessing basic foods was reported by 56.9% of women. Women were more likely to be multiparous (54.2%), have an unplanned pregnancy (61.6%), and a normal BMI (18.5–24.9 kg/m²) during early pregnancy (45.6%). Although 2.0% of women reported tobacco smoking during pregnancy while 5.3% of them reported consuming alcohol during the current pregnancy (Table 1).

Table 1:

Sociodemographic, reproductive, and psychiatric characteristics of study population

Characteristics	All participants (N = 2,051)		No stress-related sleep disturbance (N = 1,370)		Stress-related sleep disturbance (N = 681)		P-value
	n	%	n	%	n	%
Demographics
Age (years) a	27.92 ± 6.28		27.78 ± 6.32		28.21 ± 6.21		0.148
Age (years)
18–19	130	6.3	94	6.9	36	5.3	0.094
20–29	1140	55.6	755	55.1	385	56.5
30–34	429	20.9	300	21.9	129	18.9
≥35	352	17.2	221	16.1	131	19.2
Education (years)
≤6	26	1.3	18	1.3	8	1.2	0.045
7–12	871	42.6	608	44.5	263	38.8
>12	1146	56.1	740	54.2	406	60.0
Mestizo ethnicity	1652	80.6	1091	79.7	561	82.5	0.130
Married/living with a partner	1695	82.7	1124	82.2	571	83.8	0.342
Employed	1014	49.4	681	49.7	333	48.9	0.730
Difficulty accessing basic food	881	43.1	560	41.0	321	47.3	0.007
Gestational age at interview a	11.10 ± 3.95		11.16 ± 3.99		10.96 ± 3.87		0.279
Early pregnancy BMI (kg/m2)
<18.5	39	1.9	27	2.0	12	1.8	0.579
18.5–24.9	927	45.6	618	45.5	309	45.8
25–29.9	772	38.0	525	38.7	247	36.6
≥30	294	14.5	187	13.8	107	15.9
Nulliparous	938	45.8	642	46.9	296	43.5	0.154
Planned pregnancy	785	38.4	522	38.2	263	38.7	0.839
Tobacco smoking during pregnancy	41	2.0	27	2.0	14	2.1	0.899
Alcohol consumption during pregnancy	108	5.3	63	4.6	45	6.6	0.052
Psychiatric characteristics
PHQ-9 score a	7.00 ± 4.37		6.18 ± 3.91		8.62 ± 4.75		<0.001
Depression (PHQ-9 ≥10)	511 24.9		253 18.5		258 37.9		<0.001
GAD-7 score a	5.22 ± 3.84		4.49 ± 3.41		6.71 ± 4.18		<0.001
Generalized anxiety disorder (GAD-7 ≥7)	661 32.2		348 25.4		313 46.0		<0.001
PCL-C score a	23.70 ± 7.72		22.55 ± 6.46		26.02 ± 9.37		<0.001
PTSD (PCL-C ≥26)	633 30.9		338 24.7		295 43.3		<0.001
PQ-16 score a	4.05 ± 2.96		3.67 ± 2.81		4.79 ± 3.13		<0.001
High risk for psychosis (PQ-16 ≥6)	566 27.6		314 22.9		252 37.0		<0.001

For continuous variables, P-value was calculated using ANOVA; for categorical values, P-value was calculated using Chi-square test. Abbreviations: BMI, Body Mass Index; PHQ-9, Patient Health Questionnaire-9; GAD-7, Generalized anxiety disorder questionnaire-7; PTSD, posttraumatic stress disorder; PCL-C, PTSD Checklist – Civilian Version (PCL-C); PQ-16, Prodromal Questionnaire

^aMean ± standard deviation (SD)

Stress-related sleep disturbances were reported by 33.2% of women. The prevalence of depression, generalized anxiety disorder, PTSD were 24.9%, 32.2%, 30.9%, respectively. Of note, 27.6% of the study population was classified as being at a high risk of psychosis. Compared to women with no stress-related sleep disturbances, those with stress-related sleep disturbances were more likely to have difficulty accessing basic food (p=0.007. Women with stress-related sleep disorders were also significantly more likely to report antepartum depression (18.5 vs. 37.9%, p<0.001), generalized anxiety disorder (25.4 vs. 46.0%, p<0.001), PTSD (24.7 vs. 43.3%, p<0.001), and high risk for psychosis (22.9 vs. 37.0%, p<0.001) compared to those without stress-related sleep disturbances (Table 1).

Stress-related sleep disturbances were positively and statistically significantly associated with each of the psychiatric disorders studied (Table 2). Compared to women with no stress-related sleep disturbances, those with stress-related sleep disturbances had 2.69-fold higher odds of antepartum depression (OR=2.69; 95%CI:2.19–3.31). After adjusting for confounders, women with stress-related sleep disturbances had 2.74-fold higher odds of antepartum depression (aOR=2.22; 95%CI: 2.22–3.38). Women with stress-related sleep disturbances had 2.50-fold higher odds of generalized anxiety disorder compared to those without stress-related sleep disturbances (OR=2.50; 95%CI: 2.06–3.03). The association remained robust after adjusting for confounders (aOR=2.48; 95%CI: 2.04–3.02). Compared to those without stress-related sleep disturbances, women with sleep disturbances had elevated risk of PTSD (aOR=2.36; 95%CI: 1.93–2.88) and high risk of psychosis (aOR=2.07; 95%CI: 1.69–2.54).

Table 2.

Association of stress related sleep disturbances with psychiatric disorders during pregnancy (N=2,051)

	No psychiatric disorder				Psychiatric disorder
	N	%	N	%	Unadjusted OR (95% CI)	Adjusted OR (95% CI) a
Depression
No stress-related sleep disturbance	1117	72.5	253	49.5	Reference	Reference
Stress-related sleep disturbance	423	27.5	258	50.5	2.69 (2.19–3.31)	2.74 (2.22–3.38)
Generalized anxiety disorder
No stress-related sleep disturbance	1022	73.5	348	52.6	Reference	Reference
Stress-related sleep disturbance	368	26.5	313	47.4	2.50 (2.06–3.03)	2.48 (2.04–3.02)
PTSD
No stress-related sleep disturbance	1032	72.8	338	53.4	Reference	Reference
Stress-related sleep disturbance	386	27.2	295	46.6	2.33 (1.92–2.84)	2.36 (1.93–2.88)
Risk for psychosis
No stress-related sleep disturbance	1056	71.1	314	55.5	Reference	Reference
Stress-related sleep disturbance	429	28.9	252	44.5	1.98 (1.62–2.41)	2.07 (1.69–2.54)

Abbreviations: OR, odds ratio; CI, confidence interval; PTSD, posttraumatic stress disorder.

^aAdjusted for age (continuous), marital status (married vs. others), education (categories), parity (nulliparous vs. others), difficulty paying for the very basics (hard vs. not very hard)

Discussion

Stress-related sleep disturbances during pregnancy were highly comorbid and positively associated with psychiatric disorders in this cohort of pregnant Peruvian women. Compared to women without stress-related sleep disturbances, those with stress-related sleep disturbances had statistically significantly higher odds of antepartum depression, generalized anxiety disorder, PTSD, and a high risk for psychosis.

To our knowledge, there is one study that evaluated sleep reactivity in relation to psychiatric disorders. Palagini et al in their study of 62 pregnant women in Italy found that the prevalence of high stress-related sleep reactivity was 46.7.³⁷ Those with high stress-related sleep reactivity (FIRST scores ≥ 16) reported higher prevalence of depression (62.0% vs 6.1%, p < 0.001) and anxiety (55.1% vs 15.1%, p < 0.030) compared to women with low sleep reactivity (FIRST scores < 16). Our study extends this observation by examining the magnitude of association between sleep reactivity and psychiatric disorders in a large cohort of pregnant women. Although we are not able to directly compare our findings with those from other pregnancy cohorts, our findings can be tentatively compared with published studies that have associations between sleep disturbances and psychiatric disorders among men and non-pregnant women. For instance, in their study of 177 insomnia patients and 161 controls, Nakajima et al (14) reported that higher scores on a Japanese-language version of the FIRST questionnaire, indicative of sleep reactivity, were statistically significantly and positively correlated with scores on the State-Trait Anxiety Inventory (r=0.45 among insomnia patients and r=0.47 among controls, p<0.01 )¹⁴. In a subsequent study of 2,645 Japanese government employees Nakajima and colleagues reported that individuals with high (≥ 19) FIRST scores had 1.78-fold increased odds of depression (99%CI: 1.24–2.49) as compared to those with low FIRST (< 18) scores ³⁸. While few previous studies have examined sleep reactivity, other measures of sleep reactivity have associations with psychiatric disorders. Among men and non-pregnant women, sleep disturbances are associated with depression ^39,40, anxiety ^41,42, PTSD ^42,43, and psychosis ⁴⁴. On balance, the available literature ⁴⁵ including our new findings suggest that the associations of sleep reactivity with psychiatric symptoms and disorders are robust in that they have been documented in men, pregnant and non-pregnant women across vast geographical regions, across socioeconomic spectrum of high and low-income societies and using a number of different instruments to assess sleep reactivity and vulnerability to insomnia.

However, some limitations should be considered when interpreting the results from our study. Exposure and outcome assessment measures in our study were cross-sectional and based on self-report. Consequently, reported measures of association may be subject to non-systematic recall errors and systematic non-disclosure leading to misclassification. To help decrease the likelihood of reporting errors, well-trained interviewers used standardized questionnaires to collect information from participants. Additionally, interviewers and participants were unaware of specific study hypotheses. We also used previously validated assessment instruments to measure stress-related sleep disturbances and psychiatric disorders. Self-reported questionnaires are subjective measures of sleep disturbances and psychiatric disorders, which may introduce errors, although these are inherently symptom-based disorders. Furthermore, the measures we included for sleep reactivity ²⁷, depression ^30,31, generalized anxiety disorder ³², PTSD ³³, and risk of psychosis ³⁶ have all been previously validated in our study population. Lastly, our study participants were low-income pregnant women in Lima, Peru, and may not be generalizable to other pregnant populations. Future studies are also warranted to assess if sleep reactivity is different among non-pregnant women in Peru. Our study has notable strengths, including a large sample size, and the fact that information were collected by well-trained interviewers using structured and validated questionnaires. Our study extends the available literature by assessing sleep reactivity to specifically examine sleep disruptions due to stressful situations and broadly-described periods of stress⁴⁶. Furthermore, our study is the first to examine the association between stress-related sleep disturbances and psychiatric disorders among a population of pregnant women.

Overall, high sleep reactivity during pregnancy is associated with increased odds of psychiatric disorders, including depression, generalized anxiety, PTSD, and risk for psychosis. Given that psychiatric disorders during pregnancy are associated with adverse obstetrical and perinatal outcomes, it may be useful to screen for sleep reactivity to identify at-risk individuals early in pregnancy to enhance treatment and prevention strategies. Antenatal care should include early screening and intervention for both high sleep reactivity and psychiatric disorders among pregnant women.

Highlights.

• Stress-related sleep reactivity is highly prevalent in pregnancy

• Sleep reactivity was associated with increased odds of antepartum depression, generalized anxiety, post-traumatic stress disorder and high risk for psychosis

• Inquiring about sleep reactivity during antenatal care may be beneficial for identifying and caring for women at high risk of psychiatric disorders