Study Population

The Singapore Epidemiology of Eye Diseases (SEED) study comprises 10 033 individuals from 3 large population-based cross-sectional studies of Malay (Singapore Malay Eye Study, 2004-2006; N = 3280), Indian (Singapore Indian Eye Study, 2007-2009; N = 3400), and Chinese (Singapore Chinese Eye Study, 2009-2011; N = 3353) ethnicities. All studies followed the same study protocol, were conducted in the same study center (Singapore Eye Research Institute [SERI]), and recruited adults aged 40-80 years residing in the southwestern part of Singapore through an age-stratified random sampling method. The SEED methodology and population characteristics have been published elsewhere. The studies were conducted in accordance with the Declaration of Helsinki and written informed consent was obtained from all participants. All studies received approval from the SERI Institutional Review Board (R341/34/2003 for the Malay population and R498/47/2006 for the Indian and Chinese populations). For the current analysis, we excluded individuals who were not presbyopic (see definition below, N = 37), who had near vision impairment due to ocular morbidities unrelated to presbyopia (n=1291), and/or had missing data (N = 815), leaving a total of 7890 presbyopic participants for analyses.

Assessment of Near Vision and Definition of Presbyopia

Near vision was tested unilaterally by adding increments of +0.25 diopters (D) to a participant’s best-corrected distance vision correction while having them read a near logarithm of the minimal angle of resolution (logMAR) chart (Lighthouse International, New York, New York, USA) at standard reading distance (40 cm) until no further improvement in number of lines read could be observed. Presbyopia was defined as requiring a near correction of ≥+1.00 D added to a participant’s best-corrected distance vision correction to obtain a near visual acuity (VA) of ≤N8 (equivalent to 0.2 logMAR units) on the near logMAR chart in either eye (ie, objective presbyopia). We further confirmed that all eyes meeting the above definition had a best-corrected distance VA ≤0.2 logMAR units, hence minimizing the possibility of near vision impairment (VI) due to nonpresbyopic causes. Presbyopic individuals were then further categorized as corrected and uncorrected based on self-reported use of near correction (ie, reading glasses, bifocals, or multifocal glasses) obtained from standardized questionnaires.

Visual Functioning

VF was assessed with the VF-11, a modified version of VF-14 that has been changed to suit the local cultural context. The administration and Rasch validation of the VF-11 in an Asian population has been comprehensively described previously. Briefly, 11 VF questions were used to assess the level of difficulty in performing the following activities: seeing stairs, seeing street or shop signs, recognizing people, watching television, cooking, playing cards (or mahjong), reading newspapers, completing lottery forms, reading small print, driving in the day, and driving at night. The first 9 items are rated on a numeric scale ranging from 0 “no difficulty” to 4 “unable to perform activity” and the 2 driving items are rated on a 3-point scale ranging from 0 “no difficulty” to 3 “a great deal of difficulty.”

Psychometric Assessment of the VF-11

Rasch analysis was undertaken to determine the validity and measurement characteristics of the VF-11 using Winsteps software (version 3.91.2; Chicago, Illinois, USA) and the Andrich rating scale model. In brief, Rasch analysis is a form of item response theory, where ordinal ratings of the questionnaire are transformed into estimates of measures on an interval scale in logits, which improves measurement precision and limits measurement “noise.” Rasch analysis also provides extensive insight into the psychometric properties of the scale, including response category functioning, measurement precision, item “fit” to the underlying construct (eg, visual functioning), unidimensionality (ie, measurement of a single construct), targeting of item difficulty to subjects’ ability, and differential item functioning (DIF)/item bias. Rasch analysis is important for studies using rating scales, as loss of measurement quality owing to participants’ poor understanding of questions or underutilization of response categories can reduce the value of clinical research.

During Rasch analysis, coding of the VF-11 was reversed so that a higher score indicates that a person possesses better visual functioning and vice versa. The VF-11 demonstrated satisfactory fit to the Rasch model, with ordered thresholds, good range-based precision, no evidence of multidimensionality, and no DIF. However, item 6 “ playing games such as chess or cards” displayed substantial misfit (infit mean square = 1.89) and had to be removed. Following this, item 11 “driving at night” misfit slightly (1.37); however, removal of this item did not improve other fit parameters and it was therefore retained. Targeting of the composite VF-11 score was suboptimal (difference between person and item means 3.99 logits, meaning that the participants’ mean VF levels were higher than what was required to complete the VF tasks), which is not unexpected in a population-based sample where most participants were not visually impaired. In addition to the composite VF-11 score, we generated transformed individual person scores for each of the 10 items remaining after Rasch analysis.

Assessment of Other Covariates

An interviewer-administered questionnaire, standardized across all 3 studies, was used to obtain information on sociodemographic and lifestyle factors. Ethnicities were defined by the Singapore census and as indicated on the National Registration Identity Card. Participants were given the choice to be interviewed in English, Chinese, Malay, or Tamil. Variables collected include age (years), sex, ethnicity (Malay/Indian/Chinese), smoking (current/ex-smoker or nonsmoker), alcohol consumption (yes/no), highest education levels attained (<6 years/≥6 years), income (<SGD1000/≥SGD1000), and medical history (self-reported history of angina, myocardial infarction, hypertension, hyperlipidemia, thyroid problems, stroke, and diabetes).

Clinical covariates were obtained via a standardized clinical examination. Presenting distance VA was checked monocularly with participant wearing current refractive correction (if any) using a logMAR number chart at a distance of 4 meters under standard lighting by a trained optometrist. Subjective refraction, followed by assessment of best-corrected distance and near VA, was then conducted. Undercorrected refractive error was defined as an improvement of ≥2 lines on the logMAR chart after subjective refraction. Myopia was defined as spherical equivalent (SE) <-0.50 D, hyperopia was defined as SE >+1.00 D, and emmetropia as SE −0.5 to +1.00 D.

A comprehensive eye examination was conducted by an ophthalmologist, including assessment of pupillary reaction and slit-lamp biomicroscopy (Haag-Streit model BQ-900; Haag-Streit, Bern, Switzerland) for anterior segment abnormalities; measurement of intraocular pressure (IOP) with a Goldmann applanation tonometer (Haag-Streit); and gonioscopy with a Goldmann 2-mirror lens (Ocular Instruments, Inc, Bellevue, Washington, USA) under standard dark illumination for angles of the anterior chamber. All participants with IOP >21 mm Hg or narrow anterior chamber angles (temporal peripheral Van Herick grade 2 or less) underwent static automated perimetry (Swedish Interactive Threshold Algorithm standard 24-2, Humphrey Field Analyzer II; Carl Zeiss Meditec, Dublin, California, USA). A detailed examination of the lens, vitreous, and posterior segment was then conducted after pupillary dilation with 1% tropicamide and 2.5% phenylephrine hydrochloride. Two-field fundus photographs were taken using a Canon DGI nonmydriatic fundus camera and graded by trained graders at SERI for presence of any retinal abnormalities. Presence of ocular morbidities was determined by the ophthalmologist based on standardized definitions; glaucoma was diagnosed and classified using the International Society of Geographical and Epidemiological Ophthalmology Scheme, based on gonioscopy, optic disc characteristics, and visual fields results. Age-related macular degeneration (AMD) was graded from retinal photographs according to the Wisconsin Age-related Maculopathy grading system. Diabetic retinopathy (DR) was graded from retinal photographs according to a modification of the Airlie House classification system as used in the Early Treatment Diabetic Retinopathy Study.

Systolic and diastolic blood pressure (BP) was measured twice using a digital automatic blood pressure monitor (Dinamap Pro Series DP110X-RW; GE Medical Systems Information Technologies, Inc, Buckinghamshire, United Kingdom), with the average value for each parameter used in analyses. A third measurement was obtained if the 2 previous systolic BP readings differed by more than 10 mm Hg or if diastolic BP readings differed by more than 5 mm Hg. Blood samples were collected for biochemistry analysis (HbA1c, random glucose, total and low-density lipoprotein- and high-density lipoprotein-cholesterol), and DNA extraction. Based on biochemistry analyses and BP measurements, diabetes mellitus was defined as having a random glucose level of at least 200 mg/dL (to convert glucose level to millimoles per liter, multiply by 0.0555); hypertension as having a systolic BP of ≥140 mm Hg or a diastolic BP of ≥90 mm Hg; and hyperlipidemia as a total cholesterol level of at least 239 mg/dL (to convert cholesterol level to millimoles per liter, multiply by 0.0259).

Statistical Analyses

All analyses were conducted using Stata 12 for Windows (StataCorp LC, College Station, Texas, USA). To evaluate the associations of clinical and sociodemographic characteristics with uncorrected presbyopia, participants’ characteristics with and without near correction were first compared using χ ² statistic for proportions and a t test/Mann-Whitney U test for means or median as appropriate. Age-sex and multivariable logistic regression models adjusted for variables found to be significantly different between the 2 groups of participants ( Table 1 ), as well as confounders of uncorrected presbyopia reported in previous research, were used to determine the independent determinants of uncorrected presbyopia in our sample population. These confounders include age, sex, ethnicity, smoking status, history of systemic diseases (angina, myocardial infarction, thyroid problems, hypertension, hyperlipidemia, stroke, hypertension, diabetes), presence of eye conditions (undercorrected refractive error, glaucoma, AMD, cataract, DR), education, income, and presence of presenting VI in either eye. Interactions were checked for age, sex, income status, education, and ethnicity, with a P value <.1 deemed statistically significant.

To assess the impact of uncorrected presbyopia on VF, linear regression models were first used to determine the unadjusted associations of uncorrected presbyopia with the composite VF-11 score. The independent associations of uncorrected presbyopia on overall VF and individual visual tasks were determined using multiple linear regression models adjusted for age, sex, ethnicity, education, income, presenting VA in the better eye (in log units), presence of eye diseases, and comorbidities. These confounders were found to be associated with VF in unadjusted analyses ( Table 3 ) and/or have been associated with VF in previous research conducted by our group in this population.

Because most myopic individuals may have objective, but not functional, presbyopia (defined as an improvement of ≥1 line on the near logMAR chart with near correction added to presenting distance visual correction), owing to the mitigating effect of myopia on uncorrected presbyopia (as explained in the introduction), we conducted the same analyses on myopic participants only to evaluate if there was a difference in determinants and impact of uncorrected presbyopia as compared to individuals without myopia.