| Abstract: | AimTo present and evaluate a new screening protocol for amblyopia in preschool children.MethodsZagreb Amblyopia Preschool Screening (ZAPS) study protocol performed screening for amblyopia by near and distance visual acuity (VA) testing of 15 648 children aged 48-54 months attending kindergartens in the City of Zagreb County between September 2011 and June 2014 using Lea Symbols in lines test. If VA in either eye was >0.1 logMAR, the child was re-tested, if failed at re-test, the child was referred to comprehensive eye examination at the Eye Clinic.Results78.04% of children passed the screening test. Estimated prevalence of amblyopia was 8.08%. Testability, sensitivity, and specificity of the ZAPS study protocol were 99.19%, 100.00%, and 96.68% respectively.ConclusionThe ZAPS study used the most discriminative VA test with optotypes in lines as they do not underestimate amblyopia. The estimated prevalence of amblyopia was considerably higher than reported elsewhere. To the best of our knowledge, the ZAPS study protocol reached the highest sensitivity and specificity when evaluating diagnostic accuracy of VA tests for screening. The pass level defined at ≤0.1 logMAR for 4-year-old children, using Lea Symbols in lines missed no amblyopia cases, advocating that both near and distance VA testing should be performed when screening for amblyopia.Vision disorders in children represent important public health concern as they are acknowledged to be the leading cause of handicapping conditions in childhood (1). Amblyopia, a loss of visual acuity (VA) in one or both eyes (2) not immediately restored by refractive correction (3), is the most prevalent vision disorder in preschool population (4). The estimated prevalence of amblyopia among preschool children varies from 0.3% (4) to 5% (5). In addition, consequences of amblyopia include reduced contrast sensitivity and/or positional disorder (6). It develops due to abnormal binocular interaction and foveal pattern vision deprivation or a combination of both factors during a sensitive period of visual cortex development (7). Traversing through adulthood, it stands for the leading cause of monocular blindness in the 20-70 year age group (8). The main characteristic of amblyopia is crowding or spatial interference, referring to better VA when single optotypes are used compared to a line of optotypes, where objects surrounding the target object deliver a jumbled percept (9-12). Acuity is limited by letter size, crowding is limited by spacing, not size (12).Since amblyopia is predominantly defined as subnormal VA, a reliable instrument for detecting amblyopia is VA testing (13-15). Moreover, VA testing detects 97% of all ocular anomalies (13). The gold standard for diagnosing amblyopia is complete ophthalmological examination (4). There is a large body of evidence supporting the rationale for screening, as early treatment of amblyopia during the child’s first 5-7 years of life (8) is highly effective in habilitation of VA, while the treatment itself is among the most cost-effective interventions in ophthalmology (16). Preschool vision screening meets all the World Health Organization’s criteria for evaluation of screening programs (17). Literature search identified no studies reporting unhealthy and damaging effects of screening. The gold standard for screening for amblyopia has not been established (4). There is a large variety of screening methodologies and inconsistent protocols for referral of positives to complete ophthalmological examination. Lack of information on the validity (18,19) and accuracy (4) of such protocols probably intensifies the debate on determining the most effective method of vision screening (8,20-29). The unique definition of amblyopia accepted for research has not reached a consensus (4,5,30,31), further challenging the standardization of the screening protocols.Overall, two groups of screening methods exist: the traditional approach determines VA using VA tests, while the alternative approach identifies amblyogenic factors (27) based on photoscreening or automated refraction. The major difference between the two is that VA-based testing detects amblyopia directly, providing an explicit measure of visual function, while the latter, seeking for and determining only the level of refractive status does not evaluate visual function. In addition, the diagnosis and treatment of amblyopia is governed by the level of VA. On the other hand, amblyogenic factors represent risk factors for amblyopia to evolve. There are two major pitfalls in screening for amblyogenic factors. First, there is a lack of uniform cut-off values for referral and second, not all amblyogenic factors progress to amblyopia (19).Besides the issue of what should be detected, amblyopia or amblyogenic factors, a question is raised about who should be screened. Among literate children, both 3- and 4- year-old children can be reliably examined. However, 3-year-old children achieved testability rate of about 80% and positive predictive rate of 58% compared to >90% and 75%, respectively in the 4-year-old group (32). In addition, over-referrals are more common among 3-year-old children (32). These data determine the age of 4 years as the optimum age to screen for amblyopia. Hence, testability is a relevant contributor in designating the optimal screening test.If VA is to be tested in children, accepted standard tests should be used, with well-defined age-specific VA threshold determining normal monocular VA. For VA testing of preschool children Lea Symbols (33) and HOTV charts (22,32) are acknowledged as the best practice (34), while tumbling E (28,35,36) and Landolt C (28,37-39) are not appropriate as discernment of right-left laterality is still not a fully established skill (34,40). The Allen picture test is not standardized (34,41). Both Lea Symbols and HOTV optotypes can be presented as single optotypes, single optotypes surrounded with four flanking bars, single line of optotypes surrounded with rectangular crowding bars, or in lines of optotypes (22,33,34,41-53). The more the noise, the bigger the “crowding” effect. Isolated single optotypes without crowding overestimate VA (24), hence they are not used in clinical practice in Sweden (32). If presented in lines, which is recognized as the best composition to detect crowding, test charts can be assembled on Snellen or gold standard logMAR principle (34,42,51,54). Age-specific thresholds defining abnormal VA in preschool screening for amblyopia changed over time from <0.8 to <0.65 for four-year-old children due to overload of false positives (20).The outline of an effective screening test is conclusively demonstrated by both high sensitivity and high specificity. Vision screening tests predominately demonstrated higher specificity (4). Moreover, sensitivity evidently increased with age, whereas specificity remained evenly high (4). The criteria where to set the cut-off point if the confirmatory, diagnostic test is expensive or invasive, advocate to minimize false positives or use a cut-off point with high specificity.On the contrary, if the penalty for missing a case is high and treatment exists, the test should maximize true positives and use a cut-off point with high sensitivity (55). A screening test for amblyopia should target high sensitivity to identify children with visual impairment, while the specificity should be high enough not to put immense load on pediatric ophthalmologists (14). Complete ophthalmological examination as the diagnostic confirmatory gold standard test for amblyopia is neither invasive nor elaborate technology is needed, while the penalty for missing a case is a lifetime disability.In devising the Zagreb Amblyopia Preschool Screening (ZAPS) study protocol, we decided to use Lea Symbols in lines test and to screen preschool children aged 48-54 months to address the problems declared. Near VA testing was introduced in addition to commonly accepted distance VA testing (14,22,24,32,45,56-69) due to several reasons: first, hypermetropia is the most common refractive error in preschool children (70), hence near VA should more reliably detect the presence of hypermetropia; second, the larger the distance, the shorter the attention span is; and third, to increase the accuracy of the test.The pass cut-off level of ≤0.1 logMAR was defined because of particular arguments. Prior to 1992 Sweden used the pass cut-off level for screening of 0.8 (20). A change in the referral criteria to <0.65 for four-year-old children ensued, as many children referred did not require treatment (20). In addition, amblyopia treatment outcome of achieved VA>0.7 is considered as habilitation of normal vision (3,14). At last, the pass cut-off value ≤0.1 logMAR at four years can hardly mask serious visual problems, and even if they are present, we presume they are mild and can be successfully treated at six years when school-entry vision screening is performed. The aim of the ZAPS study is to present and evaluate new screening protocol for preschool children aged 48-54 months, established for testing near and distance VA using Lea Symbols in lines test. Furthermore, we aimed to determine the threshold of age-specific and chart-specific VA normative, testability of the ZAPS study protocol, and the prevalence of amblyopia in the City of Zagreb County. By delivering new evidence on amblyopia screening, guideline criteria defining optimal screening test for amblyopia in preschool children can be revised in favor of better visual impairment clarification. |
