The present study was based on data obtained from the Tohoku Medical Megabank Project Birth and Three Generation Cohort Study (TMM BirThree Cohort Study). Detailed information on the TMM BirThree Cohort Study has been published elsewhere [16]. Pregnant women and their families were recruited from July 2013 to March 2017 at obstetrical clinics and hospitals in Miyagi and Iwate prefectures in Japan. In total, 23,730 mother-child pairs participated in the study.

Among the 23,730 mother-child pairs, 1,608 were excluded because of withdrawal of informed consent, abortion, stillbirth, multiple pregnancies, or chromosome abnormalities. Of the remaining 22,122 mother-child pairs, 917 with missing data on sleep status were excluded, as well as 10,082 with missing data on the following variables: parity, prepregnancy body mass index (BMI), smoking during pregnancy, passive smoking during pregnancy, night shift, parental history of allergic diseases, maternal educational attainment, household income, maternal psychological distress, mode of delivery, exclusive breastfeeding, use of childcare facilities, and respiratory syncytial virus infection in children. Thus, data from 11,123 mother-child pairs were included in the present analyses (Fig. 1).

The presence or absence of sleep disorders during pregnancy was assessed using the Japanese version of the Athens Insomnia Scale (AIS) [17,18]. The AIS is a self-assessment psychometric instrument to quantify sleep difficulty based on the International Classification of Diseases, 10th edition criteria.

The AIS consists of 8 items: the first 5 pertain to difficulty in sleep initiation, awakening during the night, early morning awakening, total sleep duration, and overall quality of sleep, and the last 3 relate to well-being, overall functioning, and sleepiness during the day. Mothers responded to each item on a 4-point scale (0=no problem at all, 1=slightly problematic, 2=markedly problematic, and 3=extremely problematic), and symptoms were considered positive (1, 2, and 3) if they occurred at least 3 times a week in the past month.

The total score ranges from 0 to 24 points, with higher scores indicating more sleep-related problems. The cutoff score for AIS is set at 6. In the present study, we defined the presence of sleep disorders as having an AIS score of ≥6. Mothers responded to the AIS at the time of cohort registration, with a mean gestational age at registration of 19.6±7.6 weeks.

Children’s allergic diseases were assessed by mothers’ responses to a self-reported questionnaire, which included the following question: “Since the previous survey, have physicians diagnosed the children with the following diseases: bronchial asthma, atopic dermatitis, food allergy, or allergic conjunctivitis/allergic rhinitis/hay fever?” Mothers checked a box if their children developed any of the above allergic diseases. The combination “allergic conjunctivitis/allergic rhinitis/hay fever” was used for allergies affecting mucous membranes.

The survey questionnaire was sent when children were 1 month, 6 months, 1 year, 2 years, 3 years, 3 years and 6 months, 4 years, and 5 years of age. Children who developed any allergic disease or were lost to follow-up were censored.

Maternal psychological distress during pregnancy was assessed using the Japanese version of the Kessler 6 (K6) scale19,20) at the time of cohort registration. The K6 scale consists of 6 items (nervous, hopeless, restless or fidgety, so depressed that nothing could cheer you up, that everything was an effort, and worthless), and asks mothers how frequently they experienced symptoms of psychological distress during the past 30 days. Responses are rated on a 5-category scale (all of the time, most of the time, some of the time, a little of the time, and none of the time), with the total score ranging from 0 to 24 points; higher scores indicate greater distress. In this study, a cutoff of ≥13 points, which is employed to screen for serious mental illness, was used [21].

We selected potential confounders with reference to a previous report. Maternal age at delivery was categorized as <25, 25–29, 30–34, and ≥35 years. Data regarding parity, maternal height, prepregnancy weight, mode of delivery, and child’s sex were collected from medical records. Parity was categorized as never, once, twice, or 3 times or more. Prepregnancy BMI was calculated by dividing prepregnancy weight (kg) by the square of height (m²) and categorized as <18.5 kg/m², 18.5–24.9 kg/m², and ≥25.0 kg/ m². Mode of delivery was classified as vaginal or cesarean. Data on smoking (never, stopped before pregnancy, stopped after pregnancy, current), paternal smoking (never, stopped before pregnancy, stopped after pregnancy, current), the number of people who smoke inside the room, night shift (never, 1–3 times per month, 1–4 times per week, ≥5 days per week), and use of sleep medications during pregnancy were collected from questionnaires at the time of cohort registration. Smoking was categorized as “yes” if the response indicated “current.” Passive smoking was categorized as “yes” if paternal smoking was reported as “current” or the number of people who smoke inside the room was one or more. Night shift was categorized as “no” if the response was “never,” and “yes” if the response was anything other than “never.” The use of sleep medications was classified as “yes” if any of the following drugs were taken between pregnancy confirmation and cohort registration: benzodiazepine sedative-hypnotics, non-benzodiazepine sedative-hypnotics, anxiolytics (benzodiazepine derivatives), melatonin receptor agonist, and orexin receptor antagonist. Data on household income (<4.00, 4.00–5.99, ≥6.00 million Japanese yen/yr), maternal educational attainment (high school or lower, junior or vocational college, university or higher), and parental history of allergic diseases were collected from questionnaires during pregnancy and after delivery. Parental history of allergic diseases was assessed based on whether they had been diagnosed by a physician with any of the following diseases: bronchial asthma, atopic dermatitis, food allergy, allergic conjunctivitis, or allergic rhinitis. Data on exclusive breastfeeding, pet ownership of a dog or cat, use of childcare facilities, and respiratory syncytial virus infection during the first 6 months of life were obtained from questionnaires completed up to the child’s age of 6 months.

The characteristics of participants were described according to the presence or absence of sleep disorders, presenting categorical variables as counts and percentages and continuous variables as medians with interquartile ranges. We compared the incidence of each allergic disease during pregnancy according to the presence or absence of sleep disorders using Kaplan-Meier curves by performing log-rank tests. The Cox proportional hazards model was used to assess the association between sleep disorders during pregnancy and allergic diseases in children, and HRs with 95% CIs were calculated. We calculated adjusted with hazard ratios (HRs) with 95% confidence intervals (CIs) adjusted for maternal age at delivery, parity, prepregnancy BMI, child’s sex, smoking during pregnancy, passive smoking during pregnancy, night shift, use of sleep medications, maternal history of allergic diseases, paternal history of allergic diseases, maternal educational attainment, and household income (adjusted model I). We further included postnatal factors: mode of delivery, exclusive breastfeeding, pet ownership of a dog or cat, use of childcare facilities, and respiratory syncytial virus infection (only for bronchial asthma) (adjusted model II). In addition to adjusted model II, we adjusted for maternal psychological distress that significantly affected sleep as assessed by the K6 scale, taking into consideration the impact of sleep disorders due to psychological distress (adjusted model III). Moreover, for parity and psychological distress (K6 scale), which were considered particularly influential in the association between maternal sleep disorders and the development of allergic diseases in children, stratified analyses were conducted (Supplementary Tables 1, 2).

We hypothesized that sleep disorders may induce a certain level of stress, which may mediate the relationship between maternal sleep disorders and the onset of allergic diseases in children. We conducted a mediation analysis to evaluate the extent to which psychological distress during pregnancy mediated this association. We used a method based on the approach of the counterfactual framework [22,23]. The total effect can always be broken down into a natural direct effect and a natural indirect effect. The proportion of mediation was calculated as the estimated indirect effect size divided by the estimated total effect size. The detailed explanation and coding are described in previous studies [22,23].

All statistical analyses were performed using R, v.4.0.2 (R Foundation, Austria). Two-sided P<0.05 was considered statistically significant.

The TMM BirThree Cohort Study protocol was reviewed and approved by the Ethics Committee of Tohoku University Tohoku Medical Megabank Organization (2013-1-103-1; latest revision 2024-4-021). All methods were carried out in accordance with the Declaration of Helsinki. All participants provided their informed consent at enrollment. For participants with insufficient ability to understand the study protocol at any age, with the Ethics Committee’s approval, informed consent was obtained from their guardians.