Physical Activity, Sedentary Behaviors, and Retinal Vascular Caliber in Children and Adolescents With Type 1 Diabetes

Abstract

Purpose 

The aim of this study was to investigate associations of physical activity and sedentary behaviors with retinal vascular caliber in children and adolescents with type 1 diabetes.

Design 

This was a hospital-based cross-sectional study.

Methods 

A study of 122 children and adolescents with type 1 diabetes was conducted over an 8-month period. Self-reported physical activity time and time spent watching TV or playing computer or video games were obtained using interviewer-administered questionnaires. Retinal vascular caliber was measured by a trained grader using a standardized protocol and later summarized as central retinal artery equivalent (CRAE) and central retinal vein equivalent (CRVE) using a semiautomated computer program.

Results 

After adjusting for confounders (age, sex, ethnicity, body mass index, systolic blood pressure, HbA1c, maternal smoking status, age at which cow’s milk was introduced, and CRVE/CRAE, respectively), narrower CRAE was independently related to more time spent playing computer/video games [ExpB = −3.85; 95% confidence interval (CI), −6.41 to −1.29; P = 0.004], whereas wider CRVE was independently related to lower physical activity level (ExpB = −1.08; 95% CI, −2.01 to −0.15; P = 0.03) and more time spent playing computer/video games (ExpB = 4.72; 95% CI, 0.52–8.92; P = 0.02). Television viewing time was not associated with retinal vascular caliber after adjustment.

Conclusions 

The results of this study suggest that physical activity and sedentary behaviors in the form of “screen viewing time” are associated with retinal vessel caliber early in life. These results suggest that retinal vascular caliber may provide prognostic information beyond current traditional cardiovascular risk factors. Future longitudinal and interventional studies are warranted to evaluate the relevance of these observations.

Physical activity is an established modifiable risk factor for the development of cardiac morbidity and mortality.¹ Large reductions in cardiovascular disease (CVD) risk (30% to 50%) have been reported to occur in response to physical activity.² Furthermore, it has been suggested that physical activity and CVD risk factors track from childhood and adolescence into adulthood.^3,4 Currently, very little is known about the relationship between physical activity and retinal microvasculature.

Emerging evidence suggests quantitative assessment of retinal vascular caliber may have the potential to provide additional prognostic information beyond current traditional risk factors for several cardiovascular-, cerebrovascular-, and metabolic-related diseases.^5–8 Strong associations have been found between wider retinal venular caliber and narrower retinal arteriolar caliber and the incidence of vascular abnormalities in cerebral microcirculation⁵ and coronary heart disease.⁷ Furthermore, a definitive link between narrower retinal arteriolar caliber and hypertension development⁹ and wider retinal venular caliber and the increased risk of incident stroke⁶ has been reported in the literature.

The major trends highlighted in diabetes literature suggest that retinal arteriolar dilation is present in early-onset diabetes and incident diabetic retinopathy,^10–16 whereas retinal venular dilation is known to represent a later sign of diabetic retinopathy.^17–25

Tikellis and coworkers²⁶ examined the association between physical activity level, measured during leisure, sports, and work, and retinal vascular caliber in a group of older adults and found that higher levels of physical activity were associated with narrower venular caliber.²⁶ Anuradha and coworkers¹ on the other hand, who examined a younger cohort (>25 years), found that the level of physical activity was not associated with retinal vascular caliber but did note that higher amounts of sedentary behavior measured by television viewing time were associated with larger venular caliber in men.¹ Only 1 study to date has investigated the relationship among physical activity, sedentary behaviors, and retinal vascular caliber in children. Gopinath and coworkers²⁷ utilized a large nondiabetic cohort of Australian school children aged 6 years and reported that higher levels of physical activity were associated with wider retinal arterioles, whereas increased screen viewing time was associated with narrower retinal arterioles. These findings suggest that some level of relationship may exist between activity levels and retinal microvasculature.

As individuals with type 1 diabetes have a significantly higher lifetime risk of developing microvascular- and macrovascular-related diseases, research into associations between physical activity and sedentary behaviors and retinal microvasculature during childhood and adolescence is of interest to inform the development of future interventions aimed at decreasing complications of diabetes.²⁸

In this study, we examined the associations of physical activity, TV viewing time, and time spent playing computer and video games with retinal vascular caliber in a cohort of Australian children and adolescents with type 1 diabetes. We hypothesized that more time spent engaging in sedentary behaviors and lower levels of physical activity would be associated with alterations in retinal vascular caliber.

MATERIALS AND METHODS

Study Population

All participants in this study were outpatients recruited from the Department of Endocrinology and Diabetes of the Royal Children’s Hospital in Melbourne, Australia, for type 1 diabetes care. Any child or adolescent who attended the Diabetes Complications Screening Clinic for retinal imaging between July 2013 and April 2014 was eligible for inclusion in the study. During this period, a total of 122 eligible persons, aged 7 to 17 years, attended the Diabetes Complications Screening Clinic for retinal imaging. This study was approved by the human research ethics committee of the Royal Children’s Hospital (HREC–33019B) on July 4, 2013.

Sedentary Behaviors and Physical Activity Time Measures

The Australian Child and Adolescent Eating Survey (ACAES) was used to record “total screen time,” a measure of sedentary behavior. Total screen time was calculated as the time spent watching TV and/or spent playing computer and/or video games. Four categories of each measure were derived (measured in hours per day: 0–1, 2–3, 4–5, and ≥6).

An additional questionnaire was developed by the research team for the collection of physical activity information. Participants self-reported the frequency and duration (hours) of physical activity at school, organized sports, and leisure time during the previous week. To optimize the accuracy of reports, physical activity questions were completed by the participant’s parent or guardian in consultation with the participant. These questions were adapted from the physical activity questionnaire of Baecke et al.²⁹

Assessment of Sociodemographic and Clinical Variables

The maternal sociodemographic variables of marital status, level of education, country of birth, and prepregnancy age, height, and weight were ascertained via a questionnaire developed by the research team. Relevant questions relating to the mother’s pregnancy duration, health, and infant feeding habits in the first 12 months were also raised in this questionnaire.

Prenatal, perinatal, and postnatal data, collected via an audit of the participant’s medical file, included date of birth, sex, socioeconomic indexes for areas, country of birth, age at diabetes diagnosis, body mass index (BMI), HbA1c, systolic and diastolic blood pressures, biochemical data, and lipid profile information. Body mass index was calculated using height and weight measurements using standard formulae (kilograms per square meter). HbA1c was measured in a central laboratory by the high-performance liquid chromatography method. Nonfasting lipid profiles (cholesterol, triglyceride, high-density lipoprotein cholesterol, and low-density lipoprotein) were collected via standard enzymatic methods. Blood pressure was measured in the supine and standing positions using the Philips Intellivue MP5SC (Philips Medical Systems, Boeblingen, Germany).

Assessment of Retinal Vascular Caliber

Forty-five degree, nonmydriatic retinal images centered on the optic disc and macula were taken using the Canon CR6-45NM ophthalmic digital imaging system and Canon EOS 10D digital camera. Retinal images were deemed gradable if at least 1.5 disc diameters of retina nasal and temporal to the optic disc were identifiable. All 122 children and adolescents who attended the screening clinic had retinal images that were gradable for retinal vascular caliber on this basis.

Vascular caliber measurements were performed using the semiautomated computer program IVAN (Interactive Vessel Assessment), developed by the Fundus Reading Center at the University of Wisconsin, and a standardized protocol as described elsewhere.²⁴ A trained grader (S.K.) measured vascular caliber before any clinical data was collected to ensure that he was masked to any participant characteristics. Intragrader reproducibility was established by regrading 30 randomly chosen photographs [intraclass correlation coefficients of 0.95 and 0.98 for central retinal artery equivalent (CRAE) and central retinal vein equivalent (CRVE), respectively]. Measurements were summarized into CRVE and CRAE using formulae initially developed by Hubbard et al³⁰ and later revised by Knudtson et al.³¹ Arterio-to-venous ratio (AVR) was calculated as CRAE divided by CRVE. Diabetic retinopathy status was assessed from the participant’s color retinal images by an ophthalmologist masked to the participant’s clinical characteristics and was graded according to the Wisconsin level classification system.³²

Statistical Methods

All statistical analysis was performed using SPSS version 20 software. A P value of 0.05 was used for significance testing. Retinal vascular calibers were examined as continuous variables.

The univariate statistical method utilized was dependent on whether variables were normally or abnormally distributed. The parametric tests of Pearson product–moment correlation coefficient and independent samples t test were utilized for the univariate analysis of variables that were normally distributed. The nonparametric tests of Spearman rank order correlation, Mann-Whitney U, and Kruskal-Wallis test were utilized for the univariate analysis of variables that were abnormally distributed.

After univariate analysis, partial correlation and linear regression were used to explore the relationship among retinal vascular caliber measures and physical activity and sedentary behavior variables, while controlling for age, sex, ethnicity, and CRVE/CRAE. Models for venular diameter were adjusted for arteriolar diameter and vice versa to control for the potential confounding effects of fellow vessels on the outcomes of interest. A similar approach has been used in the literature.^33,34

A multivariable linear regression model was then constructed using significant associations (P < 0.05) of CRAE and CRVE revealed in univariate analysis. This model was adjusted for age, sex, ethnicity, BMI, systolic blood pressure, HbA1c, maternal smoking status, age at which cow’s milk was introduced, and CRVE/CRAE, respectively, when evaluating the relationship among measures of sedentary behaviors, physical activity, and retinal vascular caliber.

RESULTS

Of the 122 children and adolescents who underwent retinal imaging, 118 (97%) agreed to complete the ACAES and additional questionnaire. Of the 118 subjects who agreed to complete the questionnaire, the return rate of questionnaires was 69%, with 81 participants (162 images) being included in the final analysis. In the study population, right and left eye CRAE, CRVE, and AVR were highly correlated at a less than 0.0001 level. As such, similar to previous research, values for the 2 eyes were averaged in all analyses.^22,35

Key characteristics of participants who met the inclusion criteria and returned the study questionnaires and those who failed to return the study questionnaires are presented in Table 1. Participants who completed the questionnaires were similar to those who did not complete the questionnaires with respect to age, duration of diabetes, serum cholesterol, HbA1c, and BMI. However, participants who failed to return the study questionnaires were more likely to be of nonwhite ethnicity (P = 0.016). Diabetic retinopathy was not diagnosed in any participant.

TABLE 1: Key Baseline and Demographic Characteristics of Participants and Those Who Failed to Return Study Questionnaires

Univariate analysis of the relationship among sociodemographics, clinical variables, and retinal vascular caliber revealed that narrower CRAE and CRVE were related to higher lying systolic blood pressure, whereas wider CRVE was related to higher BMI, higher HbA1c, and age at which lumpy/textured food was introduced. Participants who were exposed to cow’s milk in the first 12 months of life displayed narrower CRVE and CRAE than those who were not. Furthermore, those participants whose mothers smoked during the last trimester of pregnancy displayed narrower CRAE and CRVE compared with participants whose mothers did not smoke during the last trimester. Mother’s BMI, participant’s birth weight, age at which solid food was introduced, duration of breast feeding, presence of gestational diabetes, maternal education level, and pregnancy age and duration were not related to retinal vascular caliber measures in crude analysis.

Crude analysis of measures of sedentary behaviors, physical activity, and retinal vascular caliber revealed that CRVE was inversely related to physical activity, whereas CRAE was found to be narrower in participants who spent more time playing computer or video games. A lower AVR was also related to more time spent playing computer or video games.

After adjustment for age, sex, ethnicity, and CRVE/CRAE, a moderate inverse partial correlation between CRVE and physical activity level remained (r = −0.29, P = 0.04). More time spent playing computer or video games remained related to narrower CRAE [ExpB = −4.39; 95% confidence interval (CI), −6.97 to −1.81; P = 0.001] and lower AVR (ExpB = −0.02; 95% CI, −0.03 to −0.01; P = 0.001), whereas a new association between wider CRVE and more time spent playing computer or video games was revealed (ExpB = 5.45; 95% CI, 1.30–9.60; P = 0.01) after adjustment.

In multivariable linear regression controlling for all significant associations (P < 0.05) of CRAE and CRVE revealed in univariate analysis (Table 2), narrower CRAE was independently related to more time spent playing computer/video games, whereas wider CRVE was independently related to lower physical activity levels and more time spent playing computer/video games. Television viewing time was not associated with retinal vascular caliber after adjustment.

TABLE 2: Associations Between Screen Viewing Time, Physical Activity, and Retinal Vascular Caliber

DISCUSSION

To the best of the researchers’ knowledge, this is the only study to date that has investigated the relationship among physical activity, sedentary behaviors, and retinal vascular caliber in children and adolescents with type 1 diabetes. The results of the current study suggest that physical activity and sedentary behaviors in the form of “screen viewing time” are associated with retinal microvasculature early in life.

The findings from this study show that physical activity is inversely related to retinal venular caliber in children and adolescents with type 1 diabetes. Furthermore, those who spend more time in sedentary behaviors display an adverse microvascular profile (narrower retinal arteriolar caliber and wider venular caliber).

Only 1 study to date has investigated the relationship among physical activity, sedentary behaviors, and retinal vascular caliber in children. Gopinath and coworkers²⁷ assessed a large nondiabetic cohort of Australian school children aged 6 years and reported that higher levels of physical activity were associated with wider retinal arterioles, whereas more screen viewing time was associated with narrower retinal arterioles. Similar to the study by Gopinath et al, the findings of the current study suggest that more time spent in sedentary behaviors, in the form of TV viewing and playing computer/video games, is associated with narrower arteriolar caliber.¹⁴

Contrary to the findings of Gopinath et al,²⁷ the present study reports that retinal venular caliber, not arteriolar caliber, was related to physical activity levels. In particular, wider retinal venular caliber was associated with lower physical activity levels. This finding is consistent with recent research conducted on adult populations demonstrating that higher levels of physical activity during work and organized sports were associated with a lower prevalence of diabetic retinopathy and narrower retinal venular caliber.²⁶

As alluded to previously, physical inactivity has often been related to weight gain, higher blood pressure levels, and poorer lipid profiles.³⁶ This is consistent with the findings in the present study from regression analysis that suggest a risk factor profile including more time spent in sedentary behaviors, higher systolic blood pressure, higher BMI, and lower high-density lipoprotein cholesterol predicts an increased likelihood of both retinal arteriolar narrowing and venular widening. Therefore, the researchers hypothesize that physical inactivity and sedentary behavior are related to a poorer retinal microvascular profile (narrower arterioles and wider venules), and this association may also be linked to other cardiovascular risk factors.

The findings from the present study provide the first evidence for the association between an adverse microvascular profile (narrower arterioles, wider venules) and physical inactivity and sedentary behavior early in the course of type 1 diabetes. Similar to the research by Gopinath and coworkers,²⁷ this study highlights that the association between an adverse retinal microvascular profile and poor lifestyle choices may occur as early as other established CVD risk factors³⁷ and therefore may provide supporting evidence for early intervention.

Physical activity and sedentary behaviors were assessed by self-report in this study. Therefore, it is possible that a more objective assessment of these factors might have influenced associations with retinal vascular caliber. Furthermore, although best efforts were made in this study to control for all confounders of retinal vascular caliber, the influence of other unmeasured lifestyle, genetic, and social confounders cannot be ruled out. Lastly, as this study utilized a cross-sectional design, causality cannot be inferred from the observed associations.

In summary, these findings suggest that the influence of lifestyle factors (physical activity and sedentary behaviors) on microcirculation may emerge early in life. Moreover, these results suggest that retinal vascular caliber may provide prognostic information beyond current traditional cardiovascular risk factors. However, before any significant conclusions can be drawn, future longitudinal and interventional studies are necessary to determine if early initiation of exercise and reduced screen time programs may be useful early intervention approaches for reducing the risk of both microvascular and macrovascular complications of diabetes.

ACKNOWLEDGMENTS

The authors would like to acknowledge the support of staff within the diabetes clinic at the Royal Children’s Hospital, Melbourne, Victoria, Australia, during data collection for this project. Furthermore, the authors would like to thank the grading team at the Centre for Eye Research Australia for providing retinal vascular caliber grading training for the student researcher of this project.