Epidemiolgical Studies Of Myopia In Shanghai Children And Relevant Methods For Myopia Prediction

ObjectivesTo present prevalence and incidence of myopia in Shanghai,to explore risk factors for myopia and to establish methods for myopia prediction.MethodsStratified random sampling was used to select 6 primary schools in Shanghai Baoshan district.Among them,children of grade 1 to 3 were included in the 2-year-cohort and children of grade 1 were included in the 4-year-cohort.The children underwent ocular examinations such as uncorrected visual acuity(UCVA),presenting visual acuity(PVA),axial length,slit lamp,cycloplegic autorefraction,subjective refraction,etc.Myopia risk factor related questionnaires were collected at baseline.Logistic and multiple linear regression analyses were performed to explore risk factors for myopia.Receiver-operating characteristic(ROC)curves were plotted to compare the accuracy of prediction by different predictors and the combination of them.In addition,children from another five randomly selected primary schools underwent UCVA,noncycloplegic autorefraction(NCAR),cycloplegic autorefraction,subjective refraction,etc.ROC curves were plotted to explore the best cut-off point,sensitivity,specificity,positive and negative predictive values and Youden index for UCVA,NCAR alone and the combination of the two in serial order.ResultsPrevalence of myopia increased from 1.8% in the 3-year-olds to 52.2% in 10-year-olds.The 2-year-incidence of myopia was 36.2% and mean 2-year-progression of refraction was 0.97±0.49 D towards myopia;The 4-year-incidence of myopia was 60.9% and mean progression of refraction was 2.11±0.53 D towards myopia.The risk factors of myopia were less hyperopic spherical equivalent refraction(SER)of-0.5D to 1.0D(compared with SER>1.0D),key schools(compared with normal schools),and two myopic parents(compared with no myopic parent).Baseline SER alone could effectively predict myopia incidence in 2 and 4 years with areas under the ROC curves of 0.851 and 0.836.The best cut-off to predict myopia incidence in 2 years was SER?0.5D,and the best cut-off to predict myopia incidence in 4 years was SER?0.75.Combination of UCVA?20/20 and NCAR ?-0.75 D in serial order could estimate SER of children with sensitivity and specificity of 84.4% and 90.5%.ConclusionsPrevalence of myopia is low in preschool children,however increasing rapidly after 6 years of age.The incidence of myopia is high and the progression of refraction is rapid towards myopic direction in primary school children in Shanghai.Baseline SER could be used as a single predictor to predict myopia incidence for children with relatively high accuracy and a simple procedure.Combining UCVA and NCAR could estimate children's refraction with relatively high accuracy.