Participants and the Toyama Birth Cohort Study
The participants were children from the Toyama Birth Cohort Study, a prospective, longitudinal survey examining lifestyle and health in 10,438 children born in Toyama Prefecture, Japan, between April 2, 1989 and April 1, 1990. The cohort of school children had been evaluated via a questionnaire every 3 years from Phase 1 in 1992 to Phase 5 in 2005. The overall purpose of the Toyama Birth Cohort Study was to clarify the effects of lifestyle factors and family environment on children’s health; the details of the study have been published elsewhere [7, 9-12]. The prefecture education authorities and the institutional review board at Toyama Medical and Pharmaceutical University (current University of Toyama) approved the content and ethical aspects of the current study. Written informed consent was obtained from the participants’ parents, and participation was voluntary. This study was carried out in accordance with the principles of the Declaration of Helsinki.
Questionnaire at baseline
Phase 3 of the Toyama Birth Cohort Study was conducted in 1999, when the cohort of children were in grade 4 in elementary schools (aged 9–10 years) [7, 9] . We defined Phase 3 as the baseline survey. A self-administered structured questionnaire was distributed to school children and included items pertaining to the following four main areas: lifestyle, including physical activity and food consumption frequency; psychological status; child-parent interaction; and health status, including anthropometric data and bowel movements.(Additional file 1) Children responded to questions related to their lifestyle, psychological status, child-parent interaction at home, and bowel movements, while their parents provided information regarding their children’s food consumption frequency and anthropometric data. The lifestyle items included breakfast (having every day, sometimes, or hardly), physical activity (very often, often, or rare), and hours spent watching TV on weekdays (<2, <3, or ≥3 h). Then, the answer on breakfast was dichotomized into “every day” and “skipping.” Children who reported that they “very often” or “often” did physical activity were defined as “active.” We divided sleep duration into two categories (<8 or ≥8 h), in accordance with research indicating that Japanese elementary school children’s average sleep duration was 8.5 h per night [13]. We defined <8 h sleep a day in grade 4 children as “short sleep.” With respect to food consumption frequency, we asked parents the following question: “Except school lunch, how many times does your child consume fruit and vegetables per week?” There were three response options for the question: almost every day, 3–5 times/week, or 0–2 times/week. The validity of the lifestyle questionnaire was verified in previous studies, in which frequent physical activity was significantly correlated with increased energy expenditure involving physical activity and mean steps per day (p<0.05 in a linear trend test) measured via the Actiwatch-L (Mini Mitter Company, Inc., Bend, OR) [14]. The correlation coefficient for the association between subjective and objective records of assumed sleep duration was 0.97 (p<0.001) [15] . Body mass index (BMI) was calculated using data regarding weight and height measured by trained school nurses at the children’s schools and obtained from the questionnaire. Age- and sex-specific cutoff points equivalent to the adult BMI value of 25 kg/m2 for classification as overweight were used to identify children who were overweight [16]. These cutoff points (19.46 kg/m2 for boys and 19.45 kg/m2 for girls) were developed by the Childhood Obesity Working Group of the International Obesity Task Force, and children with BMI values exceeding the cutoff points were classified as overweight.
The following questions were used to examine the children’s psychological status and parent-children’s relationship: “How often do you feel irritated?”, “How often do you feel like you do not want to go to school?”, and “How often do you usually interact with your parents?” There were four response options for the questions: often, sometimes, rarely, and none; then, the former and latter two responses were combined into a category, “not rare(ly)” and “rare(ly)” in Table 3, respectively.
Classification of constipation
The definition of constipation ranges from self-reported constipation to the fulfillment of clinical criteria involving bowel movement frequency and symptoms [1, 2]. In the current epidemiological study, we focused on only bowel movement frequency. Frequency was classified into three categories, which were as follows: at least once daily, once every 2 days, or less frequently than once every 2 days. Bowel movements less frequently than once every 2 days were defined as constipation, because this corresponds to the Rome IV criterion of “two or fewer defecations in the toilet per week” [2, 17].
Questionnaire at follow-up
In 2002, three years after the baseline survey, we followed up on the lifestyle, BMI, and bowel habits of the children in the original cohort at Phase 4 of the Toyama Birth Cohort Study, when they were adolescents in grade 7 (aged 12 to 13 years) [10]. Lifestyle at follow-up included skipping breakfast, physical activity, hours spent watching TV, and sleep duration. Sleep duration was divided into two categories (<7 or ≥7 h) on weekdays, and <7 h sleep duration was defined as “short sleep” considering children’s growth. Adolescents with BMI values exceeding the cutoff points (21.56 kg/m2 for male and 22.14 kg/m2 for female) were classified as overweight [16]. Regarding psychological status and parent-children interaction, the same three questions at baseline were used.
Changes in lifestyle, psychological status, overweight, and parent-children interaction
To clarify the importance of changes in lifestyle, psychological status, overweight, and child-parent interaction during the three years on the incidence of adolescent constipation, we divided the changes into four categories, such as “healthy to healthy,” “healthy to unhealthy,” “unhealthy to healthy,” or “unhealthy to unhealthy.”
Statistical analysis
Baseline characteristics of the distributions of lifestyle, overweight, psychological status, and children-parent interaction among the cohort of children were reported. Comparisons between adolescents with complete follow-up data and those who dropped out were conducted by chi-squared tests. Next, logistic regression analysis was performed to clarify the predisposing factors of constipation and to assess the impact of changes in lifestyle, psychological status, overweight, and children-parent interaction on constipation incidence. Crude and adjusted odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. We adopted a forced-entry method in the multivariate models. All analyses were performed using SPSS, version 25.0 J (SPSS, Chicago, IL, USA). A two-tailed p-value <0.05 was considered statistically significant.