Screen-viewing in late childhood has been associated with adiposity and blood pressure (BP), but evidence is lacking at younger ages. To investigate the prospective associations of total and device-specific screen-viewing at age 2–3 years with BMI, sum of skinfold thicknesses and BP among Singaporean children at age 3–5 years.
As part of the Growing Up in Singapore Towards healthy Outcomes (GUSTO) cohort, mothers/caregivers reported the time per day their 2 and 3-year-old children watched/used television, handheld devices and computers. Average screen-viewing time (total, television and handheld-devices) at ages 2 and 3 years was used in the analyses. Height; weight; triceps, biceps and subscapular skinfold thicknesses; and systolic and diastolic BP were measured at ages 3, 4 and 5. Associations of screen-viewing with BMI, sum of skinfold thicknesses and BP in 956 children were investigated using repeated-measures linear regression models. Analyses were further stratified by sex as we found significant interaction.
Among boys and girls combined, screen-viewing was positively associated with sum of skinfold thicknesses, but not with BMI or BP. Sex-specific analyses showed significant associations with both BMI and sum of skinfold thicknesses in boys, but not in girls. Screen-viewing was not associated with BP in boys or girls. The increases in mean (95% CI) BMI per hour increase in daily total, television and handheld-devices screen-viewing among boys were 0.12 (0.03, 0.21), 0.18 (0.06, 0.30) and 0.11 (−0.07, 0.29) kg/m2, respectively. The corresponding increases in mean sum of skinfold thicknesses were 0.68 (0.29, 1.07), 0.79 (0.26, 1.32) and 1.18 (0.38, 1.99) mm.
Greater screen-viewing at age 2–3 years was associated with later adiposity at 3–5 years in boys, but not in girls. In light of the increasing use of screen devices and cardiometabolic risk in young children, these findings may have important public health implications.
Overweight/obesity and hypertension are risk factors for adult cardiometabolic diseases including coronary heart disease and type 2 diabetes, which account for up to 7.6 million deaths per year worldwide [1, 2]. The World Health Organization has reported that 41 million children aged ≤ 5 years were overweight or obese , and international studies have reported a prevalence of hypertension/high blood pressure in children and adolescence as high as 22.3% . Overweight/obesity and elevated blood pressure (BP) are known to track from childhood to adulthood [2, 5]. Preventing excess adiposity and elevated BP in early childhood is therefore an important public health goal.
Sedentary behavior is a potentially modifiable risk factor for adiposity and BP in childhood and adolescence [6, 7]. In children, screen-viewing, including television watching and playing video or computer games, is becoming a common sedentary behavior across the world [8,9,10,11]. Previous studies have reported that 16% of Singaporean children aged ≤2 years and 25% of Canadian children aged 2–5 years spend ≥2 h per day viewing screens [9, 10]. Recent guidelines on screen time for children recommend limiting total screen time to ≤1 h per day of high-quality program for children between 2 and 5 years of age and avoiding digital media use in children younger than 18–24 months . Excessive screen-viewing has become a major public health concern, as early childhood screen-viewing behavior tracks over time .
Higher screen-viewing may be associated with greater adiposity and blood pressure in children aged 2–16 years [14,15,16,17,18,19,20,21,22]. However, the majority of previous studies have not differentiated between television viewing and other types of screen devices [20, 22,23,24,25], which are becoming increasingly common . Few studies have investigated the associations of computer, and electronic video game use with adiposity and BP in young children, and the results have been inconsistent [14, 21, 27, 28], To our knowledge, none have examined smartphones and tablets, whose usage is increasing even in toddlers [26, 29]. Furthermore, an important limitation of previous studies includes their predominantly cross-sectional design [14,15,16,17,18, 20]. Longitudinal studies are few [19, 27, 30], and to our knowledge, only one study has investigated the prospective association in children ≤5 years of age .
Evidence has also suggested that screen-viewing effects on adiposity and BP may differ by sex in older children, however, published findings remain inconsistent, warranting further investigation [23, 27, 31]. Hence, we investigated associations between screen-viewing at ages 2–3 years and adiposity and BP at ages 3–5 years, stratifying our analyses by sex.
The Growing Up in Singapore Towards healthy Outcomes (GUSTO) study is an ongoing pregnancy-offspring observational cohort. GUSTO enrolled pregnant women attending their first-trimester antenatal clinic visit between June 2009 and September 2010 at two major public maternity units in Singapore: KK Women’s and Children’s Hospital (KKH) and National University Hospital (NUH). Pregnant women <14 weeks of gestation from Chinese, Malay or Indian ethnicity, aged ≥18 years, who were Singapore citizens or permanent residents, had the intention of delivering in the study centers and of staying in Singapore for at least the next 5 years, and agreed to donate biological tissue samples were invited to participate in the study. Women with type 1 diabetes or who received chemotherapy or psychotropic drugs were excluded. The protocol of GUSTO study has been detailed previously . The study received approval from ethics committees: SingHealth Centralized Institutional Review Board and National Healthcare Group Domain Specific Review Board in Singapore. All participants gave written informed consent.
Screen-viewing time (SVT)
During clinic visits at ages 2 and 3-years, trained interviewers asked the parents or caregivers how much time their children spent in activities involving screens. Time was reported in 5-minute increments for both weekdays and weekend days. Three types of screen media were considered: 1) television viewing/playing time (e.g., PlayStation®, Wii™, Xbox™), 2) computer time (desktop or laptop), and 3) time using hand-held devices (e.g., Game Boy®, hand-phones or tablets). The average weekday and weekend day SVTs were summed, then divided by 7 [(weekday × 5 + weekend day × 2)/7] for each device-specific screen time to obtain daily television viewing time (TVT), handheld device viewing time (HDVT) and computer time. All types of screen time (TVT + HDVT + computer) were summed to obtain total SVT. To reduce reporting error, the average SVT (total, TVT and HDVT) at ages 2 and 3 was calculated and used for analyses .
Standardized protocols were used for all anthropometric measurements, as detailed previously [33, 34]. Briefly, weight (to the nearest gram) and height (to the nearest 0.1 cm) of children were measured up to three times at clinic visits at ages 2, 3, 4 and 5 years by trained research staff using a weighing scale (SECA model 803) and a stadiometer (SECA model 213, Hamburg, Germany), respectively. Triceps, biceps and subscapular skinfolds (to the nearest 0.2 mm) of the children were also measured up to five times using Holtain skinfold calipers (Holtain Ltd, Crymych, UK) at the 2, 3, 4 and 5 year clinic visits. Repeated readings of each measurement at each visit was averaged. BMI (kg/m2) was derived from average weight (in kg) divided by squared average height (in m²) for each visit. The sum of skinfolds (SSF) at each visit was calculated by adding the average biceps, triceps and subscapular skinfolds. We used BMI and SSF as measures of adiposity.
Blood pressure measurements
Blood pressure was measured at ages 3, 4 and 5 years by trained staff during clinic visits. The child was required to sit quietly with the mother for at least 5 min prior to measurement. Systolic BP (SBP) and diastolic BP (DBP) were each measured twice from the right brachial artery with the arm resting at the chest level, using a Dynamap CARESCAPE V100 (GE Healthcare, Milwaukee, WI). If there was a difference between the two readings of SBP or DBP of ≥10 mm Hg, a third measure was obtained. BPs measured when the child was crying or moving were excluded. The average of all included readings in each visit was calculated and used for analysis.
Maternal age, ethnicity, educational level and pre-pregnancy weight were obtained at recruitment as part of an interviewer-administered questionnaire. Data on children’s time spent outside in playing/exercising activities were collected with SVT data as a part of an interviewer-administered questionnaire at ages 2 and 3 years, and the average of these activities was calculated and used for analyses. Maternal self-reported pre-pregnancy weight and height measured at 26–28 weeks of gestation, and paternal weight and height were measured at the 2 or 3-year clinic visit, using a weighing scale (SECA model 803) and a stadiometer (SECA model 213, Hamburg, Germany); these were used to calculate maternal pre-pregnancy and paternal BMIs (kg/m2).
Frequencies, percentages, means and standard deviations were calculated for categorical and continuous variables, respectively. Differences in screen time (total SVT, TVT and HDVT) between groups of each characteristics of participants were tested with one-way ANOVA. Chi-square tests and independent t-tests were used to test differences between included and excluded children. Approximately 36.3% of participants had missing values for one or more of the covariates. We performed multiple imputation of 50 datasets using chained equations to handle missing data on covariates . Outcome and exposure variables were not imputed, but they were accounted in the imputation procedure . Results from these 50 datasets were combined using the multiple imputation model to provide pooled estimates. The imputation was limited to the number of children with non-missing data for the exposure and at least one of the outcomes (n = 956). Complete case analysis was carried out as a sensitivity analysis.
Multivariable repeated-measures linear regression models were used to estimate associations of average total and device-specific SVT at age 2–3 years with BMI, SSF and BP at age 3–5 years. These models account for the non-independence of repeated outcome measures at ages 3, 4 and 5 years. All models were adjusted for the following demographic and lifestyle factors as potential confounders: ethnicity, sex, average of time spent outside playing/exercising activities at age 2–3 years; maternal age, education, and pre-pregnancy BMI; paternal BMI; and study centre. In addition, BMI at age 2 years was adjusted for when assessing the associations with BMI and BP, and SSF at age 2 was adjusted to assess associations with SSF. Ethnicity, sex, maternal education and study center were analyzed as categorical variables in all models, with other covariates analyzed as continuous variables. In models with TVT and HDVT as exposures, each exposure was adjusted for the other. An interaction term between each SVT exposure and age at outcome assessment was added to the models, and tested the magnitude of associations by age. Previous studies suggest that effects of SVT on adiposity and BP may differ by sex; [37,38,39] the interaction between each SVT exposure and sex of the children was therefore tested and analyses were stratified by sex. Non-linearity of relationships between exposures and outcomes were assessed by including squared terms for SVT in models. All statistical analyses were carried out using SPSS v25 (IBM, Chicago, IL, USA).
A total of 1172 children were enrolled in the GUSTO cohort. Of those, 956 (81.6%) reported SVT and had BMI and/or sum of skinfold thickness and/or BP measured at one or more follow-up ages (Fig. 1). The 216 (18.4%) children with missing data for exposure and/or all outcome variables were similar to included children in relation to sex, ethnicity, BMI and SSF at age 2 years, and SVT (p > 0.05). SVT (total, TVT and HDVT) in study children is presented in Table 1; SVT did not vary by sex of the children (p > 0.05). However, at least one of the SVT exposures varied by ethnicity, maternal age, maternal education, maternal pre-pregnancy BMI, paternal BMI, time spent in outdoor playing/exercising activities, BMI and SSF at age 2 years and study center (all p < 0.05).
Associations of total-SVT with BMI, SSF and BP
In the overall sample, total-SVT was not associated with BMI and BP after adjusting for potential confounders [BMI: 0.04 (−0.02, 0.10) kg/m2, SBP: 0.09 (−0.24, 0.42) and DBP: 0.01 (−0.22, 0.25) mmHg]. However, total-SVT was positively associated with SSF in the adjusted model; the mean increase in SSF was 0.41 (0.13, 0.68) mm per hour increase in daily total SVT. We observed interactions between total-SVT and sex for both BMI and SSF (p < 0.10). Analysis stratified by sex showed mean increases in BMI of 0.12 (0.03, 0.21) kg/m2 and in SSF of 0.68 (0.29, 1.07) mm, for each hour increase in daily total SVT in boys. These associations were not observed in girls. Total SVT was not associated with BP in either boys or girls (Table 2).
Associations of TVT with BMI, SSF, and BP
In the overall sample, TVT was not associated with BMI after adjustment for potential confounders [0.07 (−0.01, 0.16) kg/m2]. TVT was positively associated with SSF after adjustment for potential confounders: an increase in SSF of 0.49 (0.11, 0.87) mm per hour increase in daily TVT. Interactions between TVT and sex in relation to SSF and SBP were observed (p < 0.10). Stratified analysis revealed that significant associations of TVT with BMI and SSF were observed only in boys; the mean increase in BMI per hour increase in daily TVT was 0.18 (0.06, 0.30) kg/m2, and in SSF was 0.79 (0.26, 1.32) mm. TVT was not associated with BP in overall or sex-stratified analyses (Table 2).
Associations of HDVT with BMI, SSF, and BP
HDVT was not associated with BMI or BP. In the overall sample, HDVT was associated with SSF only in adjusted analyses; the mean increase in SSF was 0.65 (0.09, 1.22) mm per hour increase in daily HDVT. Interactions between HDVT and sex in relation to BMI, SSF, SBP and DBP were observed (p < 0.10). HDVT was associated with SSF in boys [1.18 (0.38, 1.99)] but not in girls [0.14 (−0.65, 0.92)].
Further analyses revealed that the mean increases in SSF per hour increase in daily total SVT, TVT and HDVT were higher at ages 4 and 5 years than at age 3 years in the overall sample. The similar associations were found in boys but not in girls (Fig. 2). The above findings were similar in the complete case analysis (Supplementary Table 1). The test of non-linearity of the relationships between SVT and BMI and SSF was not rejected (Pnon-linearity > 0.05).
In our multi-ethnic longitudinal cohort, we observed that greater SVT was associated with higher SSF and BMI in boys, but not in girls. No associations were observed with BP in boys or girls. The findings were consistent across different types of screens that we examined. To our knowledge, GUSTO is the first study to investigate associations of total-SVT and HDVT with adiposity (BMI and SSF) and BP in children ≤5 years of age. Previously, the EDEN mother-child cohort among French children suggested that TV/DVD time of 2-year-old was not related to BMI at age 5, but related to percentage of body fat in boys . A recent review and meta-analysis of prospective studies of children of all ages reported strong evidence of a positive relationship between SVT and later BMI. The review found, however, insufficient evidence with regards to the effects of types of devices on BMI, SSF or BP .
In general, our findings in boys support previously observed associations from cross-sectional studies with predominantly older children, strengthening the evidence that greater SVT early in life, particularly TVT, is associated with higher BMI and SSF [14, 15, 21]. We found little evidence for such associations in girls. A prospective study of 6-year-old Australian children reported that increased total-SVT and TVT were associated with higher BP 5 years later . We observed no similar associations in our cohort. Differences in assessment tools, ages of the children at which exposure and outcome were measured, and duration of follow-up across studies may help explain these discrepant results.
One of the possible mechanisms underlying associations between SVT and adiposity is that SVT may displace physical activity and thus reduce energy expenditure. SVT might also increase adiposity by disrupting sleep (sleep onset latency, frequent night waking or difficulty in awakening). Finally, children who spend more time viewing screens are more exposed to food advertisements and may therefore have less favorable dietary habits, such as snacking while watching television, which could also increase adiposity [22, 24, 41,42,43].
We observed sex-specific associations in our cohort. Previous prospective studies have investigated effect modification by sex in children, but their results have been inconsistent [30, 31, 44]. A study conducted among American children (aged 9 to 16 years) found SVT to be associated with higher BMI in both sex . Another study conducted among UK children (aged 7–11 years) found TV/DVD time to be associated with increased risk of overweight in girls only . In contrast, a study among French children (aged 2 years) reported that it was associated with higher percentage of body fat in boys only . We observed stronger associations in boys. In our study context, SVT may be a better proxy for sedentary behavior in boys than in girls. For example, boys who are not watching screen devices might instead engage in physical activity, whereas girls might engage in other sedentary activities. Previous studies in older children have reported that girls have many types of sedentary behaviors and are more sedentary than boys . Further investigation in various contexts is needed to confirm these sex differences. We also observed that the magnitude of associations of SVT with SSF differed by age. Compared to the mean increase in SSF per hour increase in daily SVT at age 3 years, the mean increases were higher at ages 4 and 5, particularly in boys. It will be of great interest to follow whether these associations persist, decline, or strengthen as our cohort ages.
Our study has several important strengths. First, our longitudinal design and repeated measures of exposures and outcomes help understand the temporal relationship between SVT and adiposity. Second, we used comprehensive and objective methods to quantify adiposity and BP. Limitations of our study include the fact that SVT was reported by parents or other caregivers, rather than directly observed. However, total and device-specific SVT data were collected as part of a structured interview questionnaire and administered by trained interviewers at two time points, with the average used for all analyses, which should improve the validity of our exposure measurements. SVT was not measured at 4 or 5 years, which prevents us from investigating of associations between changes in screen time from ages 3 to 5 years and adiposity and BP. Residual confounding due to other unmeasured factors, such as moderate and vigorous physical activity, age at adiposity rebounding, BP at 2 years of age and comorbidities, cannot be excluded. Further follow-up of our cohort, including contextual information about SVT, should shed light on the relationships between SVT and later adiposity and BP. Finally, our cohort is not necessarily representative of the general Singaporean population, limiting external validity of our findings. For instance, the GUSTO cohort was designed to oversample Malay women . Nonetheless, our findings add to the existing evidence base and should be useful in designing future observational studies, as well as for developing and testing interventions to reduce early childhood SVT as part of strategies to reduce later obesity and cardiometabolic risk.
In conclusion, this study shows that total SVT, TVT, and HDVT at 2–3 years were positively associated with adiposity at 3–5 years in boys, but not in girls. No associations with BP were observed in either sex. Future studies should include both subjective and objective assessments of type and context-specific SVT and adiposity and cardio-metabolic outcomes. Use of screen devices is increasing sharply among very young children, and rigorous evidence is required to develop effective public health strategies in early childhood to prevent later adverse cardio-metabolic outcomes.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This research is supported by the Singapore National Research Foundation under its Translational and Clinical Research (TCR) Flagship Programme and administered by the Singapore Ministry of Health’s National Medical Research Council (NMRC), Singapore- NMRC/TCR/004-NUS/2008; NMRC/TCR/012-NUHS/2014. Additional funding is provided by the Singapore Institute for Clinical Sciences, Agency for Science Technology and Research (A*STAR), Singapore. KMG is supported by the National Institute for Health Research through the NIHR Southampton Biomedical Research Centre and by the European Union’s Erasmus + Capacity-Building ENeASEA Project and Seventh Framework Programme (FP7/2007-2013), projects EarlyNutrition and ODIN under grant agreement numbers 289346 and 613977. We would like to thank GUSTO study group, operational managers, research fellows, study coordinators and data management team. We greatly appreciate voluntary participation of all participants, and cooperation of KK Women’s and Children’s Hospital and National University Hospital. The GUSTO study group includes Allan Sheppard, Amutha Chinnadurai, Anne Eng Neo Goh, Anne Rifkin-Graboi, Anqi Qiu, Arijit Biswas, Bee Wah Lee, Birit F.P. Broekman, Boon Long Quah, Borys Shuter, Chai Kiat Chng, Cheryl Ngo, Choon Looi Bong, Christiani Jeyakumar Henry, Claudia Chi, Cornelia Yin Ing Chee, Yam Thiam Daniel Goh, Doris Fok, E Shyong Tai, Elaine Tham, Elaine Quah Phaik Ling, Evelyn Xiu Ling Loo, George Seow Heong Yeo, Helen Chen, Heng Hao Tan, Hugo P S van Bever, Iliana Magiati, Inez Bik Yun Wong, Ivy Yee-Man Lau, Izzuddin Bin Mohd Aris, Jeevesh Kapur, Jenny L. Richmond, Jerry Kok Yen Chan, Joanna D. Holbrook, Joanne Yoong, Joao N. Ferreira, Jonathan Tze Liang Choo, Joshua J. Gooley, Krishnamoorthy Niduvaje, Kuan Jin Lee, Leher Singh, Lieng Hsi Ling, Lin Lin Su, Ling-Wei Chen, Lourdes Mary Daniel, Marielle V. Fortier, Mark Hanson, Mary Foong-Fong Chong, Mary Rauff, Mei Chien Chua, Melvin Khee-Shing Leow, Michael Meaney, Mya Thway Tint, Neerja Karnani, Ngee Lek, Oon Hoe Teoh, P. C. Wong, Paulin Tay Straughan, Pratibha Agarwal, Queenie Ling Jun Li, Rob M. van Dam, Salome A. Rebello, S. Sendhil Velan, Seng Bin Ang, Shang Chee Chong, Sharon Ng, Shiao-Yng Chan, Shu-E Soh, Sok Bee Lim, Stella Tsotsi, Chin-Ying Stephen Hsu, Sue Anne Toh, Swee Chye Quek, Victor Samuel Rajadurai, Walter Stunkel, Wayne Cutfield, Wee Meng Han, Wei Wei Pang, Yin Bun Cheung, Yiong Huak Chan.