Prevalence of Refractive Error in Vientiane Province, Lao People’s Democratic Republic

ABSTRACT Purpose To determine the prevalence of adult refractive error and associated risk factors in the Vientiane Province of the Lao People’s Democratic Republic. Methods Population-based, cross-sectional ophthalmic survey of individuals ≥ 40 years of age in Vientiane Province. Suitable refractive data was available in 1058 individuals. Demographic data, which included age and gender, was obtained from all participants. Smoking status, presence of diabetes and hypertension was also recorded. The ophthalmic examination included autorefraction, grading of cataract, applanation tonometry and ultrasound pachymetry for ocular biometry, including axial length. Results Mean refractive error measured −0.36 diopters (D) (standard deviation [SD], 1.41) and mean cylindrical error measured −0.33 D (SD 0.95). Myopia ≤ −0.5 D and ≤ −5.0 D occurred in 53.2% (95% confidence interval [CI]: 43.7 to 62.6) and 2.0% (95% CI: 0.4 to 3.6) of participants, respectively. There was a correlation between myopia and both age and higher grading of nuclear cataract (p < .001). Hyperopia ≥ +0.5 D was present in 26.4% of participants and was associated with increasing age (p < .001). Astigmatism was present in 55.8% (95% CI: 51.5 to 60.2) of the population and was associated with increased nuclear cataract (p < .001). Urban participants had a reduced prevalence of myopia compared with rural participants. Conclusion Myopia was associated with younger age and a higher grade of nuclear cataract. The prevalence of myopia in the adult population of Vientiane Province was higher than that reported in neighbouring Asian regions and contributed to low vision.


Introduction
Although correction of refractive error is a straightforward and effective form of visual restoration, 1 uncorrected refractive error remains a leading cause of global visual impairment in developing countries. 1 This causes a significant burden at both the individual and societal economic level. 2 The prevalence of myopia varies considerably amongst ethnic groups and is increasing in certain East Asian populations due to a complex mix of genetic, environmental and socioeconomic factors. 3 The reported prevalence of myopia in South-East Asia is reported as higher than the rest of the world. [4][5][6][7] Studies from the United States report the prevalence as low as 10.4% in populations with predominantly European ancestry 8 and as high as 35% in those with Chinese ancestry. 9 The implications of uncorrected visual impairment include limiting the education and employment opportunities of economically active individuals and contributing to preventable falls in the elderly. [10][11][12] Research into the pathogenesis of myopia remains incomplete, but near-work in childhood has emerged as an important risk factor for myopia in young adults, 13 which has already led to policy changes affecting school children in some regions. [14][15][16] In older adult populations, nuclear cataract (NC) is an important risk factor for myopia. 17 Although the knowledge base regarding the prevalence of myopia and other refractive errors in Asia has increased in recent years, many ethnic and geographic gaps remain.
The Lao People's Democratic Republic (Lao PDR) is a landlocked country in South East Asia and has one of the most rapidly urbanising populations in the world. 18 Vientiane is the capital city and the associated Province comprises urban and rural areas. Optometric care is almost non-existent. 19 While there has been a recent study on the prevalence of refractive error in Vientiane school children, 20 no previous adult ophthalmic epidemiological studies have been conducted in Lao PDR. This study was prepared as part of the broader Vientiane Eye Study (VES) and is the first study to report on refractive error in the adult population of Lao PDR.

Materials and methods
The VES was a population-based, cross-sectional ophthalmic survey of inhabitants ≥ 40 years of age in urban and rural Vientiane Province. It was conducted during the period 2016-2018. According to the Lao PDR Statistics Bureau, in 2015, Vientiane Province contained approximately 12.5% of Lao PDR's population of 6.5 million and comprised rural and urban regions. 21 It is administratively subdivided into districts and villages.
The principle aim of the project was to estimate the prevalence and causes of visual impairment, including refractive error. Visual impairment was defined as presenting visual acuity worse than 6/12 in the better eye, as per World Health Organization (WHO) guidelines. 22 Participants were selected using a randomised, stratified, cluster sampling process. A sampling frame consisting of a list of all villages in the Vientiane Province along with their populations was obtained from the Ministry of Health. The sample size for the Vientiane Eye Study was calculated based on the primary end point of estimating the prevalence of blindness and low vision rather than refractive error. A sample size of 1610 was determined for a precision of 20% with 95% confidence interval (C. I.) on the a priori estimate of blindness and low vision combined of 13% (based on data in neighbouring regions), 23 and a conservatively estimated design effect of 2.0 and participation rate of 80%.
The primary sampling unit was at the village level, stratified as rural or urban. The sampling frame comprised the 491 villages in the province of which 181 (36.9%) were categorised as urban. Four urban and four rural villages were randomly selected by a computer-generated system. Households were selected by random compact sampling and all persons in the household ≥ 40 years of age were invited to participate. Sampling in each village continued until the required sample size had been reached. After obtaining written consent in the individuals' native language, personal and demographic data from the participants were collected. A single well-trained survey team were allocated specific tasks and conducted the entire study.
Each participant then received a comprehensive vision and eye examination. Specific observations were performed by 1-2 members, limiting or eliminating interobserver variability. Autorefraction, gonioscopy and disc assessment were performed by one experienced ophthalmic team member (C.C.). Examinations were performed at the Vientiane National Ophthalmology Centre, or at participants' residence if the site's distance was prohibitive.
The participant completed a standardised medical and ophthalmic questionnaire. Unaided and pinhole visual acuity was obtained using a 6 m frontilluminated E Log MAR acuity chart. Automatic objective refraction was measured with a portable keratometer-autorefractor (Retinomax K-plus 2, Right manufacturing Co. Ltd., Tokyo, Japan). Goldmann applanation tonometry was used to measure intraocular pressure (IOP) (Haag-Streit, Koeniz, Switzerland); The anterior segment was examined using a slit lamp. The presence of previous iris ischaemia or pseudoexfoliation was recorded. Gonioscopy (static and dynamic) was performed with a Sussman goniolens. Optic disc and retinal examination were performed by an experienced ophthalmologist with a 78D lens and reference to standard disc images. The vertical cup: disc ratio (CDR) and the presence of focal notching were recorded. Axial length was measured using a scanning ultrasound. Cataract was graded using the WHO simplified grading system.
The study was approved by Royal Adelaide Hospital Ethics Committee and the Lao PDR Ministry of Health and adhered to the tenets of the Declaration of Helsinki.
Myopia was deemed present if the ametropia was ≤ −0.5 diopters (D) in either eye; hyperopia was present if the ametropia was ≥ 0.5 D in either eye if the other eye was not myopic; and astigmatism was deemed present if the respective ametropia was ≥ ±0.5 D in either eye. High myopia was defined as a refractive error ≤ −5.0 D. These definitions are comparable to previous literature including the WHO guidelines on myopia. 24 For myopia, hyperopia and astigmatism either eye was included in the analysis if it met the definition. Cataract was deemed present if occuring in at least one eye. This definition has been used in similar populationbased studies. 25,26 To obtain accurate variance estimates accounting for the probability weights and correlation in the cluster survey design, the svy command in STATA v.16 (Statacorp, College Station, Texas) 27 was used to analyse the data. Logistic regression models were fitted to determine univariate and multivariate associations between a range of risk factors and each of the refractive errorrelated conditions. Linearised variance estimation was performed and prevalence, odds ratios (ORs) and 95% CIs are presented. A p-value of 0.05 was considered statistically significant.

Results
A total of 1625 participants were sampled and 1264 (77.5%) completed the full ophthalmic examination. Suitable refractive data were available in 1058 (65.1%) individuals; 661 were from rural villages and 397 from urban villages; 206 participants were excluded. The excluded participants had pseudophakia or aphakia and/or there was difficulty recording refraction due to corneal opacity or scarring. Therefore, the included study participants were 401 males and 657 females, with a mean age of 58.3 (SD 11.1) and 55.5 (SD 10.3) respectively. The most recent census data reveals a similar urban-rural population distribution (32.9 and 67.1% respectively). 21 The male to female gender ratio in the census data was 1:1, while in this study the male to female ratio was 1:1.6. Anecdotally, this was due to invited male participants being unable to leave manual work sites to attend an eye assessment. The correlation between right and left eyes for spherical equivalent (SE) was 0.655 (p < .001). The proportions are derived taking the survey weights into consideration. The particpants' characteristics and refractive data are shown in Supplementary Table 1.
Uncorrected refractive error accounted for 40.3% of presenting visual impairment and 5.9% of presenting blindness in this population. Only 25.2% of participants wore spectacles; in rural and urban locations 10.7% and 49.4% of participants wore spectacles respectively.

Relationships between participant characteristics and risk of myopia
The predictors of myopia from multivariate analyses are shown in Table 1. Younger age was a small but significant predictor of myopia. Every one-year decrease in age increased the odds of myopia by 3% (OR = 0.97, 95% CI: 0.96 to 0.97 p < .001). In the multivariate analysis, NCs and posterior subcapsular cataracts (PSC) were independently associated with myopia. For participants with NC grade of 2 or more, the odds of myopia increased 1.93 times (95% CI 1.48 to 2.51 p < .001) compared to those with no NC. For participants with grade 1 PSC, the odds of myopia increased 3.16 times (CI: 1.45 to 6.86 p = .004) and those with a grade 2 or more PSC the odds increased 2.60 times (CI: 1.21 to 5.62 p = .015) compared to those with no PSC. Additionally, with every mm increase in axial length, the odds of myopia increased 1.22 times (CI: 1.07 to 1.38 p = .002). There were no statistically significant associations with myopia concerning gender, locality, intraocular pressure (IOP), being a smoker, the presence of diabetes or hypertension.

Relationships between participant characteristics and risk of hyperopia
The prevalence of hyperopia was 26.4% (95% CI: 18.1 to 30.3). In the multivariate analysis, age, location, and PSC grade were all independently associated with hyperopia ( Table 2). Every one-year increase in age increased the odds of hyperopia 1.04 times (95% CI: 1.02 to 1.07 p < .001). Urban participants had an increased risk of hyperopia compared with rural participants (OR = 3.23, 95% CI: 1.43 to 7.30 p = .005). For participants with grade 2 or more PSC the odds of hyperopia decreased 0.51 times (OR = 0.51, 95% CI: 0.27 to 0.97 p = .041) compared to those with no PSC.

Astigmatism
Mean cylindrical error was −0.58 D [SD 1.18]. The distribution of astigmatism, overall and by gender is shown in Figure 2. In this study population, 55.8% (95% CI: 51.5 to 60.2) of the astigmatic error was ≥ ±0.5 D. In the multivariate analysis, history of diabetes, NC and cortical cataract grade were independently associated with astigmatism. Participants with diabetes had increased odds of astigmatism compared with participants without diabetes (OR = 1.46, 95% CI: 1.113 to 1.87 p = .004). For participants with grade 2 or more NC, the odds of astigmatism increased 1.87 times (95% CI: 1.36 to 2.57 p = .001) compared to those with no NC. For participants with grade 2 or more cortical cataract, the odds of astigmatism decreased 0.55 times (OR = 0.55, 95% CI: 0.44 to 0.68 p < .001) compared to those with no cortical cataract (Table 3).

Discussion
Lao PDR is an ethnically diverse country; however, the participants in the current study predominantly identified as Lao Loum people with strong linguistic associations to Thai people and probable origins from modernday southern China. Lao is a rapidly urbanising country, with most of the development occurring in Vientiane Province. The current study contained a mix of individuals from urbanising and rural regions and to our knowledge provides the first robust data about refractive error in the adult Lao population.
Although comparisons with other studies must be made cautiously due to variable definitions of refractive error, our study was consistent with the definition in the majority of reports, defining refractive error as SE ≤ −0.5 or ≥ 0.5 D. 28 For comparisons across regions, Supplementary Table 2 shows the prevalence of refractive error in selected Asian studies. The prevalence of myopia in Vientiane Province was higher than the estimated pool prevalence (EPP) of South-East Asian countries (32.9%). 28 Prevalence of myopia in Asian countries has been reported to be higher than other parts of the world. However, there is considerable regional heterogeneity in the prevalence of myopia across Asian regions. 29 There appears to be a complex interaction between genetic and environmental influences that renders myopia more prevalent in certain Asian regions. Han Chinese ethnicity may render the eye susceptible to axial myopia, becoming manifest under certain environmental conditions, particularly associated with education. 9,16 Other factors have been consistently associated with myopia, including younger age, axial length NC, PSC and near work in childhood. Myopia tends to be associated with younger age in higher income Asian regions 29 and with NC and PSC in older populations. 4,17,[30][31][32] The Korea National Health and Nutrition Examination Survey (KNHANES) and the Meiktila Eye Study (MES), both reported the association of younger age with myopia. 30,31 The Sumatra Eye study, which comprised a rural Indonesian population, also revealed a higher prevalence of myopia with younger age. 33 Near work in childhood has been increasingly recognised as a risk factor for axial myopia. 34,35 Conversely, in rural Asian regions NC-induced myopia is prevalent. 35,36 In the current study there was a higher prevalence of myopia in the younger age group, which may reflect a greater exposure to near work in childhood in the younger generation. The association between axial length and myopia is well documented 34,37 which was also demonstrated in this study. NC is a well-described cause of myopic shift and is strongly associated with myopia in population-based studies, which this study also observed. 17,38 There is considerable inconsistency in the literature surrounding the relationship between IOP and myopia. Whilst some studies have reported an association, 30,31,39,40 this study and others have found no association. [41][42][43] There was a higher prevalence of myopia in rural areas (70%) compared to urban (30%). Location however was not an independent predictor. Nonetheless the higher prevalence in rural areas may be confounded by the degree of NC in rural regions compared to urban regions.
The prevalence of high myopia (SE ≤ −5.0 D) in our study sample was 2%. This is concordant with the prevalence found in other population-based studies in Asia which reported prevalence between 1.4 and 7%. 7,33,[43][44][45][46] However, the Sumatra Eye Study, Handan Eye Study and the Namil Eye Study all reported a prevalence of high myopia of approximately 1.0%. 33,44,47 These studies were based in rural locations and may reflect the influence of childhood education on myopia pathogenesis.
The reported prevalence of hyperopia in Asia demonstrates considerable variability amongst regions. A metaanalysis reported an EPP of hyperopia in the South East Asia and Western Pacific regions of 28% and 28.5% respectively. 28 The prevalence of hyperopia in the current study (24.2%) was lower in comparison to studies of similar age groups, with reported prevalence ranging from 30% to 59%. 4,6,7,9,48 The varying difference in prevalence across regions may be due to a negative association with cataract. In the current study the odds of hyperopia decreased by 49% in participants with a grade 2 or more PSC. Age was found to be independently associated with hyperopia. This is expected, as the ageing eye undergoes a hyperopic shift change between 31 to 64 years and then a myopic shift in older age, due to cataract. Location was also found to be associated with hyperopia. This could be explained by a younger population living in urban areas who are experiencing a hyperopic shift.
The prevalence of astigmatism (55.8%) was higher than the EPP reported in South East Asia (44.8%) and Western Pacific (44.2%). 16 In the multivariate analysis, presence of diabetes and NC grade were found to be independently associated with astigmatism. There is evidence that lenticular astigmatism is observed in diabetic patients. 49,50 This was a large, randomised study that surveyed individuals from a variety of geographical locations across Vientiane Province. Nonetheless this study does have limitations. The male to female ratio recruited for this study was lower than the recent census data, and this gender imbalance may introduce bias in the estimates. A response rate of 65.1% was not as high as other studies reporting the prevalence of refractive error. There was no information available to enable nonresponse adjustment to the sampling weights. Therefore, the validity of the prevalence's reported relies on the assumption that nonresponse is entirely random concerning the outcomes of refractive error. Additionally, the sample size calculation of this study was based on a desired precision for the estimation of visual impairment, the primary outcome of the VES, rather than an estimate for the prevalence of refractive error. This is likely to have inflated confidence intervals around the refractive error prevalence estimation. The variance estimate does not assume a normal approximation; hence, the confidence interval is not symmetric about the estimated proportion. ). In terms of the exclusion criteria, corneal opacity or scarring meant that refraction could not be obtained. This could alter the true prevalence of refractive error through selection and observer bias. There may be a possibility of systematic bias due to patients with symptomatic eye disease having difficulty attending the study but conversely could also render them more likely to attend to access medical assistance. This study provides the first robust data on the refractive status of the adult population of Vientiane Province. Myopia was prevalent in the younger adult population but was more evident in older individuals with a high degree of NC. The prevalence values found were higher than that documented in previous studies. This may be related to higher levels of untreated NC in this population.
For this reason, active programs are being implemented to increase the availability of spectacles and cataract surgery to those in the Vientiane Province. 52,53 In addition to this, eye care professionals have been trained at provincial, district and local clinic levels to deliver services to remote areas. 52 This will improve the treatments available for refractive errors which form a large proportion of blindness and low vision in the population.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
The project was financially supported by Sight For All, who empower communities to deliver comprehensive, evidencebased high quality eye health care through the provision of research, education, and equipment.

Declaration
This submission has not been published anywhere previously and it is not simultaneously being considered for any other publication.