Cohort Profile: Shahroud Schoolchildren Eye Cohort Study (SSCECS)

The World Health Organization (WHO) estimates that 253 million people live with visual impairment worldwide, for which uncorrected refractive errors and untreated cataracts are the leading causes. This is although over 80% of visual impairments are preventable and treatable.1 Today, 19 million children under the age of 15 years have visual impairments, out of which 12 million are due to refractive errors. The global eye health action plan 2014–19 aims to reduce avoidable visual impairment by 25% by 2019. It emphasizes the need for biomedical studies to develop preventive measures for visual impairment, especially in low- and middle-income countries.2 Myopia is the most common type of paediatric visual impairment3 which has been increasing since 1950 in most parts of the world, especially South East Asian countries. As such, over 90% of the teenagers in China and South Korea are myopic.4 The prevalence of myopia in other countries seems to be increasing too.5 It is estimated that, by 2015, half the world population will be myopic and 10% will have high myopia by 2050.6 The biological aetiology of myopia is still unknown, and there are no standard prevention methods3 nor theoretical consensus on its risk factors.7 As a result, it continues to remain a growing and usually unpreventable problem.3,4,6 Studies on ocular biometry have contributed to a better understanding of the aetiology of refractive errors.8–13 For example, lens power decreases with age and hyperopic shift or myopia may be the result of such changes.14 In this regard, prospective cohort studies are necessary to further our knowledge. Except for the Anyang Childhood Eye Study (which was conducted with a sufficient sample size in China in 201315 and is expected to provide evidence on paediatric visual impairments) and a study in Australia,16 other longitudinal studies on children have lacked sufficiently large sample sizes,17–20 included only one follow-up20–22 or were not population-based.11,19,20,23 Given the role of genetic and environmental factors in the incidence of visual impairment, and the fact that most myopia studies have been done in East and South East Asia, it is essential to develop and implement cohort studies in other parts of the world. Moreover, the availability of state-of-the-art ophthalmic and biometric equipment in recent years should enhance our ability to dissect the aetiology of visual impairments and ocular biometric changes. Following the successful implementation of the Shahroud Eyes Cohort Study (ShECS) in 40- to 64-year-old participants in 2009,24 a cohort study on children living in this region may offer a better understanding of the risk factors and aetiology of visual impairments among children and adults. The Shahroud Schoolchildren Eye Cohort Study (SSCECS), whose baseline data were collected from February to September 2015, was conducted with an adequate sample size and with 3-year follow-ups aimed to respond to the following questions. What are the prevalence and incidence of refractive errors, especially myopia, in children and what factors influence them? How do refractive errors change and progress over time, and what factors affect them? What are the prevalence proportion and incidence rate of other visual problems, such as amblyopia, strabismus and binocular vision disorders, among 6- to 12-year-old children, and what factors affect them? How do biometric and corneal topographic indices change during emmetropization, and what factors affect them? What is the status of oral health, especially the decayed, missing, and filled teeth (DMFT) index, among 6- to 12-year-old children, and what factors affect it? The present study was conducted in Shahroud, north-eastern Iran, and was supported by Noor Ophthalmology Research Center and Shahroud University of Medical Sciences. The children were selected from urban and rural areas of Shahroud, using a stratified sampling method. For a better estimation of the rural population and provision of better services to deprived areas, all primary schoolchildren in rural areas (a total of 1214 children) were invited to participate. In urban areas, children were selected using cluster random , and classrooms were selected as clusters. Since the average size of the 473 eligible classes was 27 students, 200 classes were selected using systematic random sampling to provide an adequate number of participants. All students in the selected classrooms were provided with a full description of the study objectives and methods, the informed consent form, and a questionnaire to collect contact information, ophthalmic records, insurance history, students’ daily activities on schooldays and weekends and their nutritional status (in the second phase). After coordinating with schools, an examination day was scheduled for each classroom, and students whose parents had signed the informed consent form were transferred to the examination site. Students whose parents did not sign the consent form were considered as non-participants and were not followed up. Overall, 6624 students were invited and 5620 students (84.8%) participated in the study; the sample included 1154 students (95.1%) of the 1214 rural students and 4466 students (82.6%) of the 5410 urban students. Of the participants 54% were male and 46% were female. Table 1 compares the demographic characteristics of the children who participated with those of non-participants. The mean age of the participants (midway in the project) was 9.77 years and the mean age of the non-participants was 9.79 years; the difference was not statistically significant (P = 0.788). The comparison of respondent and non-respondent participants, Shahroud, Iran, 2015 The comparison of respondent and non-respondent participants, Shahroud, Iran, 2015 The population living in Shahroud, for many demographic indices, is largely similar to the average in Iran, and the results of this study can be generalized to the Iranian population, albeit with caution. For example, based on the National Population and Housing Census in 2016, the percentage of literacy in the population over 6 years in Iran and Shahroud was 87.6% and 92.5%, the unemployment rate in the active population was 12.6% and 11.8%, the household size was 3.3 and 3.1, the below-14 years population was 24.0% and 21.4% and those who lived in rental homes was 30.8% and 31.6%, respectively. Similar to the average in Iran, the ethnicity of most people in Shahroud is Fars and 99.8% are Muslim.25 In the first phase of the study in 2015, 84.8% of the invited students participated in the study. The second phase of the study is still ongoing at the time of writing this cohort profile in 2018, and about 90% of the first phase participants are expected to participate in the second phase. Other phases of study will be done at 3-year intervals up to five phases. Participants will leave the cohort when they reach the age of 18 years, at some point between the third and fifth phases of study. A summary of the first and second phase measures is presented in Table 2. Further descriptions of these measurements are provided below. Summary of baseline data available in the SSCECS and percentage providing data (n = 5620), Shahroud, Iran, 2015 Summary of baseline data available in the SSCECS and percentage providing data (n = 5620), Shahroud, Iran, 2015 After obtaining the signed informed consents and recording their contact information, parents were asked a number of questions about their employment status, eye care insurance status and their children’s daily tasks and habits. Participants’ height was measured in the upright position without shoes, using a stadiometer with an accuracy of 0.1 cm. The participants were weighed with minimal clothing, using digital scales with an accuracy of 0.1 kg. The body mass index (BMI) was calculated as weight (in kilograms) divided by the square of height (in metres squared). The stadiometer and digital scales were checked daily, and calibrated if necessary. The WHO 2007 Reference and AnthroPlus software26 was used to calculate z-scores for weight-for-age (WAZ), height-for-age (HAZ) and BMI-for-age (BAZ). Waist, hip and arm circumferences, upper leg length and triceps skinfold thickness were also measured using standard methods.27 Waist circumference was measured using a non-elastic tape at mid distance between the lower edge of the last thoracic vertebra and iliac crest. Hip circumference was measured around the widest portion of the buttocks, with the tape parallel to the floor. Body composition was measured using Body Composition Analyzer (BoCA x1, Medigate Inc., Korea). After a 15-min rest break, the blood pressure and heart rate of each participant were measured and recorded twice with a 5-min interval. This was done in the sitting position using the participant’s right arm, by trained nurses using a calibrated digital barometer (OMRON, HEM-2228-E). To this end, an appropriately sized blood pressure cuff was used. Measurements were done a third time if there was a difference of more than 10 mmHg in systolic measurements or more than 5 mmHg in the diastolic blood pressure readings. The diagnosis of hypertension was based on the definition described in the Fourth Report on the Diagnosis, Evaluation, and Treatment of High Blood Pressure in Children and Adolescents,28 and after calculating HAZ. First, non-cycloplegic autorefraction was carried out using the Nidek ARK 510 A auto-refracto-keratometer. Then, uncorrected visual acuity was measured using the Nidek CP-770 chart projector at a distance of 3 m Autorefraction results were then refined using retinoscopy, and objective refraction was recorded. Finally, subjective refraction was measured, and the best corrected visual acuity was recorded. Measurement of the amplitude of accommodation and near point of convergence was performed in accordance with the procedures described elsewhere.29 Binocular stereopsis was examined using Stereo Fly test (Stereo Optical Company Inc., Chicago, USA) with reference distance constant of 15 min of Arc. The presence of tropia and phoria at near and far distances was determined by optometrists using the unilateral cover test (cover-uncover test) and the alternating cover test, respectively. Finally, iris colour was determined before cycloplegia with adequate lighting and was classified in eight categories. Ocular biometric components were measured using the Allegro Biograph (WaveLight AG, Erlangen, Germany) before cycloplegia. The output of this device was electronically exported into the study database. Tomography and anterior segment imaging were carried out using the Pentacam HR (Oculus Inc., Lynnwood, WA, USA, with the Oculus software 6.03r19/ 1.18r08). Various variables were automatically exported into the study database. These included pachymetry, posterior and anterior corneal astigmatism, corneal imaging and keratometry. After instilling two drops of cyclopentolate 1% in each eye at 5-min intervals, cycloplegic autorefraction and retinoscopy were done at least 30 min after the instillation of the second drop. If required, a third drop was instilled in the eyes. When applicable, eyeglasses were prescribed based on cycloplegic refraction results at least 3 days after the initial examination. Oral and dental examinations were carried out by two trained dentists using disposable dental mirrors, dental explorers and a dental torch. In the second phase of the study, measurement of body composition was omitted and the following items were added: imaging with optical coherence tomography (OCT), measurement of intraocular pressure with a non-contact tonometer, colour vision testing, monocular estimated method retinoscopy, and assessment of positive and negative relative accommodation. In the next phases, which are to be implemented at 3-year intervals, all examinations of the second phase will be repeated. In the first phase of the study, 5620 students participated and the following results, which are also summarized in Table 3, were obtained. The resulting publications are available at [www.sscecs.info]. The main results of the first phase of Shahroud school children eye cohort study, Shahroud, Iran, 2015 The main results of the first phase of Shahroud school children eye cohort study, Shahroud, Iran, 2015 Mean amplitude of accommodation was 14.44 dioptres (D) [95% confidence intervals (CI): 14.33–14.55 D], which decreased with age and was lower in rural schoolchildren. It was higher in myopes at 17.30 D (95% CI: 16.74–17.86 D). In all age groups, the measured amplitude of accommodation was lower than the values calculated with the Hofstetter’s equation.29 Mean near point of convergence was 8.08 cm (95% CI: 7.96–8.19 cm), which was higher than other studies and increased with age, spherical equivalent and near point of accommodation.30 Mean central corneal thickness was 556.29 µm (95% CI: 555.01–557.57 µm). Central corneal thickness was directly correlated with the mean keratometry and anterior chamber depth, and measurements were lower in girls and rural participants.31 A comparison between autorefraction and cycloplegic retinoscopy showed higher spherical error values with autorefraction. However, this difference is not clinically significant, and autorefraction can be used instead of retinoscopy.32 Mean lens power was 22.86 (95% CI: 22.75–22.98) D, which was higher among girls, rural participants and hyperopic children. Lens power decreased with age and significantly correlated with axial length, anterior chamber depth, spherical equivalent, lens thickness and corneal power.33 The prevalence of hypertension and pre-hypertension in the first phase of the study was 6.8% (95% CI: 5.9–7.8) and 7.4% (95% CI: 6.7–8.2), respectively, in total participants. In rural areas, this prevalence was 22.7% (95% CI: 18.2–25.2) and 15.0% (95% CI: 12.4–17.5), respectively. High BMI, living in rural areas, female gender and low and moderate economic status were associated with an increased risk of hypertension and pre-hypertension.34 Mean WAZ, HAZ and BAZ were 0.063, 0.004 and 0.080, respectively. The mean values of all three indices were significantly lower in rural areas than in urban areas (P <0.001). Overall, 25.7% (95% CI: 24.1–27.4) of the urban children and 14.8% (95% CI: 12.3–17.3) of the rural children were overweight or obese. A total of 5.1% (95% CI: 3.1–7.0) of the rural girls had short stature (unpublished paper). A total of 46 students [0.79% (95% CI: 0.55–1.04)] had amblyopia. The prevalence of amblyopia in urban and rural areas was 0.77% (95% CI: 0.51–1.03) and 1.05% (95% CI: 0.34-1.76), respectively (unpublished paper). The prevalence of astigmatism (cylinder power ≥0.50 D in at least one eye) was 16.7% (95% CI: 15.6–17.7); the rate was higher in urban areas [17.2% (95% CI: 16.0–18.3)] than in rural areas [12.1% (95% CI: 10.0–14.1)] (P<0.001) (unpublished paper). Optical biometry results showed that axial length was almost similar to that in European countries (Table 3) but shorter than that reported from South East Asian countries. Ocular biometric components were influenced by age, gender and height (unpublished paper). Mean keratometry was 43.6 D (95% CI 43.5–43.6). This value was 0.82 D higher in girls than in boys, and 0.62 D higher in myopic children than in emmetropic children.35 The most common corneal topography pattern in the right eye was oval [32.7% (95% CI: 30.1–34.5)] (unpublished paper). Some of the strengths of the present study were its large sample size, high participation rate in the first phase (84.8%), accurate study design and monitoring of its implementation, electronic data registration, use of artificial intelligence to prevent incorrect data registration and use of cycloplegic refraction and modern, accurate ophthalmic equipment such as Pentacam. However, the study had some weaknesses in the first phase, including the lack of slit-lamp examinations, lack of tonometry and optical coherence tomography (OCT), blood pressure measurement on only one occasion and use of a pencil for the measurement of near point of convergence. An application form is available on the cohort website [www.sscecs.info]. All national and international collaborations for using cohort data are welcome. For cohort data, researchers can also directly contact the first author (M,H,E,) by e-mail [emamian@shmu.ac.ir]. This cohort study aimed to identify the prevalence, incidence, causes and risk factors of visual impairment in children. This study, with its large samples size and high participation rate, is one of the few studies to perform cycloplegic refraction. In the first phase of the study conducted in 2015 in Shahroud, north-east Iran, 5620 students aged 6 to 12 years participated in the study. The participants were from urban and rural areas, and 54% of them were boys. The second phase of the SSCECS, started in 2018, is under way at the time of preparing this paper. Data were collected through: (i) interviews and general examinations; (ii) comprehensive optometric examinations; (iii) ocular biometry and tomography; and (iv) oral health examinations. To access data, researchers should first fill out the relevant form available at [www.sscecs.info]. Conflict of interest: None declared.