Prevalence, Correlates, and Impact of Uncorrected Presbyopia in a Multiethnic Asian Population
Total Page:16
File Type:pdf, Size:1020Kb
Prevalence, Correlates, and Impact of Uncorrected Presbyopia in a Multiethnic Asian Population RYAN EYN KIDD MAN, EVA KATIE FENWICK, CHARUMATHI SABANAYAGAM, LING-JUN LI, PREETI GUPTA, YIH-CHUNG THAM, TIEN YIN WONG, CHING-YU CHENG, AND ECOSSE LUC LAMOUREUX PURPOSE: To examine the prevalence, correlates, and and Indian ethnicities are needed. (Am J Ophthalmol impact of uncorrected presbyopia on vision-specific func- 2016;168:191–200. Ó 2016 Elsevier Inc. All rights tioning (VF) in a multiethnic Asian population. reserved.) DESIGN: Population-based cross-sectional study. METHODS: We included 7890 presbyopic subjects (3909 female; age range, 40-86 years) of Malay, Indian, RESBYOPIA, AN AGE-RELATED INABILITY TO FOCUS and Chinese ethnicities from the Singapore Epidemiology on near objects owing to a loss of accommodative 1 of Eye Disease study. Presbyopia was classified as amplitude, is believed to be extremely prevalent P 2 corrected and uncorrected based on self-reported near or nearly universal in individuals above the age of 65, correction use. VF was assessed with the VF-11 question- although recent literature has suggested that this accom- naire validated using Rasch analysis. Multivariable logis- modative loss may be complete by as early as 50 years of 3 tic and linear regression models were used to investigate age. Although not a blinding condition, presbyopia signif- the associations of sociodemographic and clinical param- icantly reduces quality of life (QoL), especially if not 4 eters with uncorrected presbyopia, and its impact on corrected. However, direct estimates of the prevalence VF, respectively. As myopia may mitigate the impact of of uncorrected presbyopia, its associated risk factors, and noncorrection, we performed a subgroup analysis on its impact on QoL are limited, particularly in Asian myopic subjects only (n [ 2742). populations. RESULTS: In total, 2678 of 7890 subjects (33.9%) had In Asia, the rates of myopia are extremely high. For uncorrected presbyopia. In multivariable models, younger instance, almost half of Singaporeans aged 40 years and 5 age, male sex, Malay and Indian ethnicities, presenting above have some degree of myopia. Uncorrected myopia distance visual impairment (any eye), and lower educa- may help mitigate the clinical impact of presbyopia, since tion and income levels were associated with higher odds the focal point of objects in myopic eyes falls further in of uncorrected presbyopia (all P < .05). Compared front of the retina as compared to nonmyopic eyes; this with corrected presbyopia, noncorrection was associated consequently results in less accommodative amplitude with worse overall VF and reduced ability to perform in- needed to bring near objects into focus. Unfortunately, dividual near and distance vision-specific tasks even after the impact, if any, of this high rate of myopia on uncorrec- adjusting for distance VA and other confounders (all P < ted presbyopia is unknown. .05). Results were very similar for myopic individuals. In this study, we investigated the prevalence, correlates, CONCLUSION: One-third of presbyopic Singaporean and impact of uncorrected presbyopia on vision-specific adults did not have near correction. Given its detrimental functioning (VF) in a multiethnic sample of Asian adults impact on both near and distance VF, public health stra- in Singapore. A secondary aim was to determine the impact tegies to increase uptake of presbyopic correction in of the high rates of myopia on the determinants and impact younger individuals, male individuals, and those of Malay of uncorrected presbyopia in this sample population. Supplemental Material available at AJO.com. Accepted for publication May 20, 2016. From the Singapore Eye Research Institute, Singapore, Singapore (R.E.K.M., E.K.F., C.S., L.-J.L., P.G., Y.-C.T., T.Y.W., C.-Y.C., E.L.L.); METHODS Centre for Eye Research Australia, University of Melbourne, Melbourne, Victoria, Australia (E.K.F., T.Y.W.); Duke-NUS Graduate STUDY POPULATION: The Singapore Epidemiology of Medical School, Singapore, Singapore (E.K.F., C.S., T.Y.W., E.L.L.); Eye Diseases (SEED) study comprises 10 033 individuals Singapore National Eye Centre, Singapore, Singapore (T.Y.W., C.-Y.C., E.L.L.); and National University of Singapore, Singapore, Singapore from 3 large population-based cross-sectional studies of (T.Y.W., C.-Y.C., E.L.L.). Malay (Singapore Malay Eye Study, 2004-2006; N ¼ Inquiries to Ecosse Luc Lamoureux, Director, Population Health, 3280), Indian (Singapore Indian Eye Study, 2007-2009; Singapore Eye Research Institute (SERI), 20 College Road, The ¼ Academia Discovery Tower Level 6, Singapore 169856; e-mail: ecosse. N 3400), and Chinese (Singapore Chinese Eye Study, [email protected] 2009-2011; N ¼ 3353) ethnicities. All studies followed 0002-9394/$36.00 © 2016 ELSEVIER INC.ALL RIGHTS RESERVED. 191 http://dx.doi.org/10.1016/j.ajo.2016.05.019 the same study protocol, were conducted in the same study PSYCHOMETRIC ASSESSMENT OF THE VF-11: Rasch anal- center (Singapore Eye Research Institute [SERI]), and ysis was undertaken to determine the validity and measure- recruited adults aged 40-80 years residing in the south- ment characteristics of the VF-11 using Winsteps software western part of Singapore through an age-stratified random (version 3.91.2; Chicago, Illinois, USA) and the Andrich sampling method. The SEED methodology and population rating scale model.13 In brief, Rasch analysis is a form of characteristics have been published elsewhere.6,7 The item response theory, where ordinal ratings of the question- studies were conducted in accordance with the naire are transformed into estimates of measures on an in- Declaration of Helsinki and written informed consent terval scale in logits, which improves measurement was obtained from all participants. All studies received precision and limits measurement ‘‘noise.’’14–16 Rasch approval from the SERI Institutional Review Board analysis also provides extensive insight into the (R341/34/2003 for the Malay population and R498/47/ psychometric properties of the scale, including response 2006 for the Indian and Chinese populations). For the category functioning, measurement precision, item ‘‘fit’’ current analysis, we excluded individuals who were not to the underlying construct (eg, visual functioning), presbyopic (see definition below, N ¼ 37), who had near unidimensionality (ie, measurement of a single vision impairment due to ocular morbidities unrelated to construct), targeting of item difficulty to subjects’ ability, presbyopia (n¼1291), and/or had missing data (N ¼ and differential item functioning (DIF)/item bias. Rasch 815), leaving a total of 7890 presbyopic participants for analysis is important for studies using rating scales, as loss analyses. of measurement quality owing to participants’ poor understanding of questions or underutilization of response 17–19 ASSESSMENT OF NEAR VISION AND DEFINITION OF categories can reduce the value of clinical research. PRESBYOPIA: Near vision was tested unilaterally by adding During Rasch analysis, coding of the VF-11 was reversed increments of þ0.25 diopters (D) to a participant’s best- so that a higher score indicates that a person possesses bet- corrected distance vision correction while having them ter visual functioning and vice versa. The VF-11 demon- read a near logarithm of the minimal angle of resolution strated satisfactory fit to the Rasch model, with ordered (logMAR) chart (Lighthouse International, New York, thresholds, good range-based precision,20 no evidence of New York, USA) at standard reading distance (40 cm) un- multidimensionality, and no DIF. However, item 6 ‘‘playing til no further improvement in number of lines read could be games such as chess or cards’’ displayed substantial misfit observed. Presbyopia was defined as requiring a near correc- (infit mean square ¼ 1.89) and had to be removed. tion of >_þ1.00 D added to a participant’s best-corrected Following this, item 11 ‘‘driving at night’’ misfit slightly distance vision correction to obtain a near visual acuity (1.37); however, removal of this item did not improve (VA) of <_N8 (equivalent to 0.2 logMAR units) on the other fit parameters and it was therefore retained. Target- near logMAR chart in either eye (ie, objective presby- ing of the composite VF-11 score was suboptimal (differ- opia).8,9 We further confirmed that all eyes meeting the ence between person and item means 3.99 logits, above definition had a best-corrected distance VA <_0.2 meaning that the participants’ mean VF levels were higher logMAR units, hence minimizing the possibility of near than what was required to complete the VF tasks), which is vision impairment (VI) due to nonpresbyopic causes. Pres- not unexpected in a population-based sample where most byopic individuals were then further categorized as participants were not visually impaired. In addition to corrected and uncorrected based on self-reported use of the composite VF-11 score, we generated transformed indi- near correction (ie, reading glasses, bifocals, or multifocal vidual person scores for each of the 10 items remaining af- glasses) obtained from standardized questionnaires. ter Rasch analysis. VISUAL FUNCTIONING: VF was assessed with the VF-11, ASSESSMENT OF OTHER COVARIATES: An interviewer- a modified version of VF-14 that has been changed to suit administered questionnaire, standardized across all 3 the local cultural context.10 The administration and Rasch studies, was used to obtain information on sociodemo- validation