Global Prevalence and Causes of Visual Impairment and Blindness in Children: A Systematic Review and Meta-Analysis

Abstract

Purpose:

To determine the global prevalence and common causes of visual impairment (VI) and blindness in children.

Methods:

In this meta-analysis, a structured search strategy was applied to search electronic databases including PubMed, Scopus, and Web of Science, as well as the list of references in the selected articles to identify all population-based cross-sectional studies that concerned the prevalence of VI and blindness in populations under 20 years of age up to January 2018, regardless of the publication date and language, gender, region of residence, or race. VI was reported based on presenting visual acuity (PVA), uncorrected visual acuity (UCVA), and best corrected visual acuity (BCVA) of equal to 20/60 or worse in the better eye. Blindness was reported as visual acuity worse than 20/400 in the better eye.

Results:

In the present study, 5711 articles were identified, and the final analyses were done on 80 articles including 769,720 people from twenty-eight different countries. The prevalence of VI based on UCVA was 7.26% (95% confidence interval [CI]: 4.34%–10.19%), PVA was 3.82% (95% CI: 2.06%–5.57%), BCVA was 1.67% (95% CI 0.97%–2.37%), and blindness was 0.17% (95% CI: 0.13%–0.21%). Refractive errors were the most common cause of VI in the subjects of selected articles (77.20% [95% CI: 73.40%–81.00%]). The prevalence of amblyopia was 7.60% (95% CI: 05.60%–09.10%) and congenital cataract was 0.60% (95% CI: 0.3%–0.9%).

Conclusion:

Despite differences in the definition of VI and blindness, based on PVA, 3.82%, and based on BCVA, 1.67% of the examined samples suffer from VI.

INTRODUCTION

Visual impairment (VI) in childhood has a negative and sometimes irreversible impact on children's psychological, educational, and social performance, which can persist into adulthood and affect individuals' quality of life.1 Given the significant burden of VI, its causes, and visual complications, the VISION 2020 Initiative was implemented by the World Health Organization (WHO) to eliminate preventable blindness on a global level.2,3 According to WHO estimates at the beginning of the VISION 2020 program, about 19 million children under the age of 15 years were visually impaired and 1.4 million children had irreversible blindness, and it was predicted that half of the blindness cases were preventable.4 The reported prevalence of blindness in low and middle-income countries ranges from 0.2 to 7.8/10,000 people, and in developed and industrialized countries, the annual incidence is 6/10,000 in the under-15 age group.5,6 According to available information, the causes of VI differ by the residence location of the studied population (urban versus rural) or in different countries (developed, under developed, or developing) as well as the prevention strategies within each health system. Nevertheless, Courtright et al. suggest that retinal disorders, glaucoma, corneal ulcers due to vitamin A deficiency, cataract, and neural causes are the most common causes of VI in low and middle-income countries.5 This is while neurological disorders are one of the major causes of VI in industrialized countries, and in countries such as England, 75% of blindness cases are due to unpreventable causes.7,8 A large amount of information on VI in children has been generated from population-based and clinic-based studies, studies in schools for blind children, different age groups (3–5 years, 7-years, 3–10 years, under-15-years, 5–15 years, etc.) as well as different settings such as high-income and low-income countries, but due to the mentioned differences, it is not possible to make global policies or evaluate measures that have been taken in this regard. Given the lack of cohesive results on the prevalence of VI as well as the differences in the causes of VI in different parts of the world, it seems necessary to have an estimate of the global prevalence and causes of VI in children to inform policies, especially the Vision 2020 Initiative. Therefore, the present study aims to determine the overall prevalence and causes of VI in children in the world.

METHODS

The entire process of this study was conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.9 All population-based cross-sectional studies concerning the prevalence of VI and blindness in individuals under 20 years of age were reviewed regardless of publication and language, gender, region of residence, and race. The search strategy and entry terms showed in Appendix 1. Of studies conducted on the same population, the one with a higher quality was included in this review. Also, we included studies that were performed in all age groups and used the prevalence rates reported for the under-20-year age groups. We excluded articles that did not have one or more of the inclusion criteria. The outcome of interest was the prevalence of VI and blindness and the causes of VI in the population. In the selected papers, cases of VI were identified using measurements based on different units including feet, logMAR, and meters. For this reason, and to facilitate the presentation of the results, all measurements were converted to feet.

The prevalence of VI in this study was calculated based on uncorrected visual acuity (UCVA), best corrected visual acuity (BCVA), and presenting visual acuity (PVA) as reported in previous studies.10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40 The participant's PVA was considered UCVA in participants without glasses and visual acuity with present glasses in individuals with glasses. According to previous studies, the prevalence of VI was reported based on visual acuity cut-point of 20/40 or worse and 20/60 or worse in the better eye (according to the WHO guidelines, VI based on PVA, UCVA, and BCVA was considered as visual acuity in the better eye of equal to 20/60 or worse). The prevalence of blindness was determined based on: (1) BCVA of 20/200 or worse in the better-seeing eye, and (2) BCVA of 20/400 or worse in the better-seeing eye (according to the WHO guidelines, blindness was defined as visual acuity worse than 20/400 in the better eye). We excluded the studies that specifically investigated the VI and blindness in the schools for the blind.

To ensure the correct selection of articles related to the topic of the research and in accordance with the inclusion criteria, two researchers (E.H. and M.S.) independently selected the articles; they were not blinded to the names of the authors, the journal titles, or study results. The kappa agreement index between researchers was 80.2%. Cases of controversy between the researchers were decided through discussion or by consulting a third person. The two researchers independently extracted the required data based on predefined variables. We used the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist41 to perform a qualitative assessment of the selected articles in terms of methodology and report. Present key elements of study design, describe the setting, locations, and relevant dates, including periods of recruitment, give the eligibility criteria, and the sources and methods of selection of participants, clearly define all outcomes, and report numbers of outcome events or summary measures were assessed. The studies were categorized as low risk of bias if they reported all items, as moderate risk of bias if they reported all items but one, and as high risk of bias otherwise. To examine the inconsistency of the articles, the k-square test was used at a 5% confidence interval (CI). In order to quantitatively analyze the heterogeneity of the results, we used the I-square test based on the Higgins classification. According to which, an I-square more than 75% was considered as heterogeneity. The variables investigated in this study included the name of the first author, the year of publication, the country of the study, the mean age and gender distribution of study subjects, sample size, the prevalence VI (based on UCVA, PVA, and BCVA) and blindness with their 95% CI, and the prevalence of the most important causes of VI and blindness. One of the PRISMA checklist items is calculating publication bias. In our study, publication bias was not assessed because the prevalence is always a positive number between zero and one, and cannot be negative; therefore, all studies were distributed on the right side of the vertical line, and this leads to asymmetry in the funnel plot which is not related to publication bias. Data analysis was performed using Stata Software version 11 (StataCorp, College Station, TX, USA). The data was analyzed using the random-effects model at a 95% confidence level.

RESULTS

In the present study, 5711 studies were identified; 5211 articles by searching electronic databases and 500 articles through the lists of references of selected articles and other sources. After removing redundant articles, the title and abstract of 4381 articles were reviewed, and 4231 articles were excluded after applying the exclusion criteria, and thus, 150 papers were eligible for full-text review. After reviewing the full text of the articles, 70 articles were excluded from the study for not meeting the inclusion criteria, lack of access to the full text of the article, nonoriginal paper (letter, commentary, review), and finally, data for this study were extracted from 80 articles [Figure 1].

Figure 1.

Flow of information through the different phases of the systematic review

As shown in Table 1, the final 80 papers comprised 769,720 people from a total of 28 different countries.10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83

Table 1.

Summary of studies results

1st author	Country (city)	Gender percentage male	Age mean, range	SS	UCVA % (95% CI)
Abu-Shagra et al., 199142	Saudi Arabia	100	10.9 (6-19)	1188	-
Adhikari et al., 201443	Nepal	47.3	5.7±3.1 (0-10)	10,950	-
Ajaiyeoba et al., 200744	Nigeria	44.1	11.8±3.8 (4-18)	1144	-
Akogun 199245	Nigeria	54.5	9-19	1600	-
A1 Faran et al., 199346	Saudi Arabia	49.0	0-19	1909	-
Alrasheed et al., 201647	Sudan	-	6-15	1678	6.40 (4.90-7.90)
Beiram 197148	Sudan	-	0-19	127,426	-
Bucher and Ijsselmuiden, 198849	South Africa	40.5	0-19	44,977	-
Casson et al., 201250	Asia	49.9	6-11	2899	-
Congdon et al., 200827	China	50.2	14.7±0.8 (11.4-17.1)	1892	41.17 (38.94-43.42)
Dandona et al., 199940	India	-	0-15	663	-
Dandona et al., 200151	India	-	0-15	2859	-
Darge et al., 201752	Ethiopia	50.8	11.05±2.5 (5-16)	378	-
Demissie and Solomon, 201153	Ethiopia	-	0-15	58,480	-
Dorairaj et al., 200854	India	-	3-15	13,241	-
Drews et al., 199255	Atlanta	-	10	89,534	-
Farber 200356	Israel	48.6	0-18	1161	-
Feghhi et al., 200957	Iran	40.5	5-19	2492	-
Flanagan et al., 200358	Ireland	-	10.5±4.8 (1-18)	47,110	-
Fotouhi et al., 200729	Iran	52.1	7-15	5544	-
Ghosh et al., 201259	India	45.8	6-14	2570	4.24 (3.41-5.10)
Gilbert et al., 200826	Six countries	51.7	5-15	40,779	-
Goh et al., 200532	Malaysia	50.8	7-15	4634	17.07 (15.99-18.18)
Hashemi et al., 201811	Iran	-	1-15	766	-
He et al., 201417	China	57.9	7-12	9512	13.33 (12.65-14.03)
He et al., 200728	China	52.5	13-17	2454	27.04 (25.27-28.86)
He et al., 200433	China	51.9	5-15	4364	22.27 (21.04-23.54)
Heijthuijsen et al., 201360	Suriname	-	8-16	4643	-
Jamali et al., 200961	Iran	-	6	902	3.55 (2.39-5.07)
Johnson and Minassian, 198962	Africa	-	0-6	5436	-
Kaphle et al., 201663	Malawi	54.8	0-15	635	-
Kedir and Girma, 201464	Ethiopia	54.1	7-15	592	-
Kemmanu et al., 201665	India	-	≤15	23,087	-
Khandekar et al., 200266	Oman	52.1	0-15	6208	-
Kingo and Ndawi, 200967	Tanzania	-	6-17	400	-
Kumah et al., 201319	Ghana	46.6	12-15	2453	3.65 (2.94-4.47)
Li et al., 201568	China	51.5	0-19	22,148	-
Limburg et al., 201220	Vietnam	52.2	0-15	28,800	-
Lu et al., 200924	Beijing	52.2	4.41±1.09 (0-6)	17,699	-
Ma et al., 201613	China	54.0	3-10	8267	19.79 (18.93-20.66)
Maul et al., 200039	Chile	50.7	5-15	5303	15.72 (14.75-16.73)
Moraes Ibrahim et al., 201318	Brazil	51.0	12.4±1.6 (10-15)	1590	5.72 (4.63-6.98)
Moser et al., 200269	Equatorial Guinea	47.9	0-19	812	-
Murthy et al., 200236	India	51.9	7-15	6447	6.40 (5.79-7.05)
Naidoo et al., 200335	South Africa	49.3	5-15	4679	1.34 (1.03-1.71)
Newland et al., 199270	Vanuatu	-	6-19	483	-
O’Donoghue et al., 201071	Northern Ireland	50.5	13.1±0.38 (12-13)	661	12.85 (10.40-15.65)
Pai et al., 201121	Sydney	51.3	2-4	475	-
Pan et al., 201612	China	53.3	4-6	713	-
Park et al., 201416	South Korea	52.6	5-19	4394	-
Paudel et al., 201415	Vietnam	46.1	12-15	2238	19.39 (17.77-21.09)
Pi et al., 201272	Western China	52.4	6-15	3079	-
Pokharel et al., 200038	Nepal	51.7	5-15	4803	2.87 (2.41-3.38)
Premsenthil et al., 201373	Malaysia	49.0	4-6	400	-
Raihan et al., 200574	Bangladesh	50.2	5-15	28,835	-
Razavi et al., 201075	Iran	-	6-13	123	-
Rezvan et al., 201276	Iran	41.5	11.2±2.4 (6-17)	1547	2.20 (1.41-2.90)
Robaei et al., 200531	Sydney	50.6	6.7 (5-9)	1738	1.32 (0.84-1.97)
Rustagi et al., 201277	Delhi	46.8	14.25 (11-18)	1075	2.88 (1.96-4.06)
Salomão et al., 200978	Brazil	48.2	11-14	2440	4.83 (4.01-5.76)
Sapkota et al., 200825	Kathmandu	53.5	10-15	4282	18.63 (17.47-19.83)
Sewunet et al., 201479	Ethiopia	43.1	7-15	420	11.66 (8.75-15.12)
Shahriari et al., 200780	Iran	46.2	10-19	2307	-
Sharma et al., 201781	Haryana	40.3	6-15	1265	2.68 (1.86-3.73)
Srivastava and Verma, 197882	India	54.4	0-14	7822	-
Tabbara and Ross-Degnan, 198683	Saudi Arabia	50.4	0-19	4467	-
Tananuvat et al., 200484	Chiang Mai	-	6-7	3467	-
Taylor et al., 201085	Australia	-	5-15	1694	-
Thulasiraj et al., 200334	India	-	6-19	5342	-
Unsal et al., 200986	Turkey	53.7	10.52±2.2 (6-17)	1606	-
Varma et al., 201787	United States	-	3-5	-	-
Vitale et al., 200630	United States	43.8	12-19	4564	-
Wu et al., 201388	China	52.9	9.7±3.3 (4-18)	6026	27.09 (25.97-28.24)
Xiao et al., 201189	China	-	<16	23,675	-
Yamamah et al., 201514	Egypt	50.6	10.7±3.1 (5-17)	2070	29.42 (27.46-31.43)
Yekta et al., 201022	Iran	53.5	10.9±2.2 (7-15)	1872	6.46 (4.96-7.96)
Zainal et al., 200290	Malaysia	47.0	0-9	4690	-
Zerihun and Mabey, 199791	Ethiopia	50.5	0-19	4084	-
Zhao et al., 200037	China	48.8	5-15	5884	12.81 (11.97-13.69)
MEPEDS Group 200923	African-American Hispanic	-	2-6	1592 165	-
Abu-Shagra et al., 199142	11.86 (10.08-13.84)	-		≤6/12 in the better eye	Medium risk
Adhikari et al., 201443	0.1 (0.04-0.15)	-	0.07 (0.02-0.12)	VI: <6/18 in the better eye BL: PVA <6/60	Low risk
Ajaiyeoba et al., 200744	1.32 (0.74-2.18)	0.17 (0.02-0.63)	VI: <6/18 either in one or both eyes BL: VA <3/60	Low risk
Akogun 199245	8.12 (6.83-9.57)	-	3.81 (2.92-4.87)	VI: <6/18 in the better eye BL: VA <6/60 in the better eye	High risk
A1 Faran et al., 199346	-	1.67 (1.14-2.35)	-	<6/18 in the better eye	Medium risk
Alrasheed et al., 201647	4.40 (2.90-5.90)	1.20 (0.30-2.70)	-	≤6/12 in the better eye	Low risk
Beiram 197148	-	-	0.071 (0.057-0.087)	VA ≤3/60 in the better eye	High risk
Bucher and Ijsselmuiden, 198849	-	-	0.006 (0.001-0.019)	PVA <3/60 in the better eye	High risk
Casson et al., 201250	1.90 (1.43-2.46)	-	-	<20/32 in the better eye	Low risk
Congdon et al., 200827	19.29 (17.53-21.14)	0.47 (0.21-0.90)	-	≤6/12 in the better eye	Low risk
Dandona et al., 199940	2.86 (1.73-4.43)	-	-	<20/40 in the better eye	Medium risk
Dandona et al., 200151	-	-	0.17 (0.05-0.40)	PVA <6/60 in the better eye	Medium risk
Darge et al., 201752	5.82 (3.68-8.67)	-	-	≤6/12 in the either eye	Low risk
Demissie and Solomon, 201153	-	-	0.05 (0.03-0.07)	PVA <6/60 in the better eye	Low risk
Dorairaj et al., 200854	-	-	0.11 (0.06-0.17)	BCVA <3/60 in the better eye	Low risk
Drews et al., 199255	-	-	0.068 (0.05-0.08)	BCVA <20/200 in the better eye	High risk
Farber 200356	-	-	14.41 (12.41-16.60)	VA ≤20/400 in the better eye	High risk
Feghhi et al., 200957	-	5.09 (4.26-6.03)	-	<20/60 in the better eye	Medium risk
Flanagan et al., 200358	-	0.057 (0.03-0.08)	-	≤6/18 in the better eye	High risk
Fotouhi et al., 200729	1.73 (1.40-2.11)	0.25 (0.13-0.42)	-	≤20/40 in the better eye	Low risk
Ghosh et al., 201259	-	0.19 (0.06-0.45)	-	<6/12 in the better eye	Medium risk
Gilbert et al., 200826	-	0.14 (0.11-0.18)	-	<6/18 in the better eye	Low risk
Goh et al., 200432	10.08 (9.22-10.98)	1.42 (1.10-1.81)	2.033 (1.64-2.48)	VI: ≤20/40 in the better eye BL: ≤20/200 in the better eye	Low risk
Hashemi et al., 201711	1.30 (0.63-2.38)	0.52 (0.14-1.33)	0.78 (0.28-1.69)	VI: ≤20/60 in the better eye BL: VA <20/400 in the better eye	Low risk
He et al., 201417	11.25 (10.63-11.91)	0.63 (0.48-0.81)	-	≤20/40 in the better eye	Low risk
He et al., 200728	16.58 (15.11-18.13)	0.45 (0.22-0.81)	-	≤20/40 in the better eye	Low risk
He et al., 200433	10.25 (9.36-11.19)	0.61 (0.41-0.89)	-	≤20/40 in the better eye	Low risk
Heijthuijsen et al., 201360	2.30 (1.89-2.77)	-	0.81 (0.57-1.12)	VI: <6/18 in the better eye BL: PVA <3/60 in the better eye	Medium risk
Jamali et al., 200961	-	-	-	<6/12 in either eye	Medium risk
Johnson and Minassian, 198962	-	-	0.11 (0.04-0.24)	VA<3/60 in the better eye	Medium risk
Kaphle et al., 201663	3.60 (0.43-12.31)	-	1.78 (0.04-9.55)	VI: VA <6/18 in the better eye BL: PVA <3/60 in the better eye	Medium risk
Kedir and Girma, 201064	1.75 (0.84- 3.20)	1.40 (0.61-2.74)	-	<6/18 in the better eye	Low risk
Kemmanu et al., 201565	-	-	0.077 (0.046-0.12)	BCVA <3/60 in the better eye	Low risk
Khandekar et al., 200266	-	-	0.08 (0.02-0.18)	PVA<3/60 in the better eye	Low risk
Kingo and Ndawi, 200967	9.50 (6.81-12.80)	-	-	VI: VA<6/18 in the better eye	Medium risk
Kumah et al., 201319	3.53 (2.83-4.34)	0.41 (0.19-0.75)	-	≤20/40 in the better eye	Low risk
Li et al., 201568	-	0.07 (0.04-0.11)	0.02 (0.007-0.05)	VI: <6/18 in the better eye BL: BCVA <3/60 in the better eye	Low risk
Limburg et al., 201220	-	-	0.07 (0.05-0.11)	PVA <3/60 in the better eye	Medium risk
Lu et al., 200924	0.42 (0.33-0.53)	-	-	<6/18 in the better eye	Medium risk
Ma et al., 201613	15.53 (14.75-16.33)	1.69 (1.42-1.99)	-	≤20/40 in the better eye	Low risk
Maul et al., 199939	14.57 (13.63-15.55)	7.29 (6.61-8.03)	-	<20/40 in at least one eye	Low risk
Moraes Ibrahim et al., 201318	2.83 (2.07-3.76)	0.81 (0.43-1.39)	-	≤20/40 in the better eye	Medium risk
Moser et al., 200269	-	-	0.61 (0.20-1.43)	VA <3/60 in the better eye	Medium risk
Murthy et al., 200136	4.85 (4.32-5.43)	0.81 (0.59-1.06)	-	<20/40 in the better eye	Low risk
Naidoo et al., 200335	1.17 (0.88-1.52)	0.32 (0.17-0.52)	-	VA ≤20/40 in the better eye	Low risk
Newland et al., 199270	-	-	0.21 (0.005-1.14)	VA <6/18 in the better eye	High risk
O’Donoghue et al., 201071	3.17 (1.97-4.81)	-	-	<6/12 in the better eye	Low risk
Pai et al., 201121	6.10 (4.12-8.65)	-	-	<20/50 in the better eye	Low risk
Pan et al., 201612	6.59 (4.88-8.66)	-	-	<20/40 in the better eye	Low risk
Park et al., 201416	6.12 (5.43-6.87)	-	0.25 (0.12-0.44)	VI: <20/60 in the better eye BL: VA <20/400 in the better eye	Low risk
Paudel et al., 201415	12.19 (10.87-13.62)	-	0.26 (0.09-0.58)	VI: VA ≤6/12 in the better eye BL: PVA ≤6/120 in the better eye	Low risk
Pi et al., 201272	7.69 (6.78-8.69)	-	-	≤20/40 in the better eye	Low risk
Pokharel et al., 200038	2.83 (2.38-3.34)	1.35 (1.04-1.72)	-	≤20/40 in the better eye	Medium risk
Premsenthil et al., 201373	5.0 (3.08-7.61)	-	-	≤6/12 in the better eye	Low risk
Raihan et al., 200574	-	-	0.06 (0.04-0.11)	PVA <3/60 in the better eye	High risk
Razavi et al., 201075	-	-	17.88 (11.56-25.81)	VA <3/60 in the better eye	Low risk
Rezvan et al., 201276	1.0 (0.59-1.67)	0.25 (0.07-0.66)	-	≤6/12 in the better eye	Low risk
Robaei et al., 200531	0.86 (0.48-1.41)	-	-	≤20/40 in the better eye	Low risk
Rustagi et al., 201277	-	-	0.93 (0.44-1.70)	VI: <20/60 in the better eye BL: VA <20/200 in the better eye	Medium risk
Salomão et al., 200978	2.70 (2.09-3.42)	0.40 (0.19-0.75)	-	≤20/40 in the better eye	Low risk
Sapkota et al., 200825	9.08 (8.24-9.98)	0.86 (0.60- 1.18)	-	≤20/40 in the better eye	Medium risk
Sewunet et al., 201479	-	6.42 (4.27-9.21)	-	<20/40 in the better eye	Medium risk
Shahriari et al., 200780	-	1.51 (0.98-2.04)	<20/60 using a pinhole	Low risk
Sharma et al., 201781	-	-	-	≤6/18 in the better eye	High risk
Srivastava and Verma, 197882	-	-	0.14 (0.07-0.25)	PVA <3/60 in the better eye	High risk
Tabbara and Ross-Degnan, 198683	11.59 (10.67-12.57)	-	2.39 (1.96-2.88)	VI: <6/18 in the better eye BL: PVA <3/60 in the better eye	Low risk
Tananuvat et al., 200484	8.68 (7.76- 9.66)	-	-	≤20/40 at least one eye	Medium risk
Taylor et al., 201085	1.68 (1.12-2.43)	-	0.18 (0.03-0.52)	VI: <6/12 in the better eye BL: PVA<6/60 in the better eye	Low risk
Thulasiraj et al., 200334	0.73 (0.52-0.99)	0.48 (0.32-0.72)	0.07 (0.02-0.19)	VI: <6/18 in the better eye BL: PVA<3/60 in the better eye	Low risk
Unsal et al., 200986	1.68 (1.11-2.43)	-	-	<20/40 in the better eye	High risk
Varma et al., 201787	1.50 (1.20-1.80)	-	-	<20/50 or 20/40 in the better eye	Low risk
Vitale et al., 200630	9.70 (8.86-10.60)	-	-	≤20/50 in the better eye	Low risk
Wu et al., 201388	-	0.31 (0.19-0.49)	-	≤20/40 in the better eye	Low risk
Xiao et al., 201189	-	-	0.02 (0.006-0.049)	PVA <3/60 in the better eye	Medium risk
Yamamah et al., 201514	-	-	-	≤6/9 in the better eye	Medium risk
Yekta et al., 201022	1.49 (0.82-2.15)	0.90 (0.30-2.74)	-	≤6/12 in the better eye	Low risk
Zainal et al., 200290	0.44 (0.27-0.68)	-	0.04 (0.005-0.15)	VI: <6/18 in the better eye BL: PVA <3/60 in the better eye	Low risk
Zerihun and Mabey, 199791	0.18 (0.04-0.53)	-	0.07 (0.01-0.21)	VI: <6/18 in the better eye BL: PVA <3/60 in the better eye	High risk
Zhao et al., 200037	10.92 (10.14-11.75)	1.75 (1.43-2.11)	-	≤20/40 in the better eye	Low risk
MEPEDS Group 200923	2.76 (2.01-3.69)	0.78 (0.41-1.33)	-	<20/50 or 20/40 in the better eye	Low risk
	2.47 (1.77-3.35)	0.71 (0.36-1.22)

SS: Sample size, UCVA: Uncorrected visual acuity, PVA: Presenting visual acuity, BCVA: Best corrected visual acuity, CI: Confidence interval, VI: Visual impairment, BL: Blindness, VA: Visual acuity

Among the selected articles, the studies by Razavi et al.75 in Iran with 123 people and Beiram84 with 127,426 people in Sudan had the smallest and the largest sample sizes, respectively.

The overall prevalence of VI was 12.72% (95% CI: 9.26%–16.19%) based on a UCVA of 20/40 or worse in the better eye, and 7.26% (95% CI: 4.34%–10.19%) based on a UCVA of 20/60 or worse in the better eye [Figure 2]. The prevalence was 7.34% (95% CI: 5.53%–9.15%) based on a PVA of 20/40 or worse in the better eye and 3.82% (95% CI: 2.06%–5.57%) with a PVA of 20/60 or worse in the better eye, and 2.91% (95% CI: 2.31%–3.51%) based on a PVA worse than 20/60 in the better eye [Figure 3]. The prevalence of VI based on a BCVA of 20/40 or worse in the better eye was 0.77% (95% CI: 0.56%–0.97%), 1.67% (95% CI 0.97%–2.37%) based on a BCVA of 20/60 or worse in the better eye, and 0.88% (95% CI: 0.63%–1.12%) based on a BCVA worse than 20/60 in the better eye [Figure 4].

Figure 2.

Overall prevalence and subgroups of uncorrected visual acuity based on uncorrected visual acuity

Figure 3.

Overall prevalence and subgroups of presenting visual acuity (PVA) based on PVA

Figure 4.

Overall prevalence and subgroups of best corrected visual acuity (BCVA) based on BCVA

Based on criteria worse than 20/200 in better eye and worse than 20/400 in the better eye, the blindness prevalence was 0.15% (95% CI: 0.06%-0.25%) and 0.17% (95% CI: 0.13%-0.21%), respectively [Figure 5]. Table 2 summarizes the prevalence of UCVA, BCVA, PVA VI, and blindness in the six regions of the WHO. The highest rate of VI based on UCVA of 20/40 or worse in the better eye was 20.10% (95% CI: 13.75%–26.45%) in the Pacific Region, and based on UCVA of <20/60 in the better eye was 15.72% (95% CI: 14.74%–16.70%) in the Americas. The highest prevalence of VI based on PVA of 20/40 or worse in the better eye, 20/60 or worse in the better eye, and worse than 20/60 in the better eye in the Pacific Region was 10.87% (95% CI: 7.26%–14.48%), 8.03% (95% CI 1.00% -20.84%) in the Americas, and 11.59 (95% CI: 10.65–12.53) in the Eastern Mediterranean Region, respectively. The highest prevalence of VI based on a BCVA of 20/40 was 0.91 (95% CI: 0.54–1.27) in the Pacific Region. The highest rates of blindness were 1.91 (95.1% CI: 1.78–5.58) in the African Region based on worse than 20/200 and 1.94 (95% CI: 0.27%–3.61%) in the Eastern Mediterranean Region with criteria worse than 20/400.

Figure 5.

Overall prevalence and subgroups of blindness

Table 2.

Prevalence of visual impairment and blindness in the six regions of the World Health Organization

WHO region	UCVA % (95% CI)		PVA % (95% CI)			BCVA % (95% CI)		Blindness % (95% CI)
	≤20/40 in better eye	≤20/60 in better eye	≤20/40 in better eye	≤20/60 in better eye	<20/60 in better eye	≤20/40 in better eye	≤20/60 in better eye	<20/200 in better eye	<20/400 in better eye
Eastern Mediterranean	4.24 (1.00-8.55)	7.47 (1.00-15.43)	3.62 (1.81-5.44)	1.54 (1.04-2.04)	11.59 (10.65-12.53)	0.41 (0.12-0.71)	3.36 (1.00-9.14)	-	1.94 (0.27-3.61)
Americas	5.19 (4.34-6.04)	15.72 (14.74-16.70)	2.75 (2.25-3.25)	8.03 (1.00-20.84)	-	0.57 (0.18-0.96)	7.29 (6.59-7.99)	0.07 (0.05-0.08)	-
Africa	3.76 (1.09-6.44)	-	3.57 (1.58-5.56)	-	3.48 (1.96-5.01)	0.55 (0.19-0.91)	0.78 (0.36-1.21)	1.91 (1.78-5.58)	0.11 (0.04-0.16)
Western Pacific	20.10 (13.75-26.45)	6.10 (3.95-8.25)	10.87 (7.26-14.48)	2.90 (1.42-4.37)	2.11 (0.97-3.23)	0.91 (0.54-1.27)	-	0.17 (0.01-0.37)	0.05 (0.02-0.08)
South-east Asia	7.77 (1.15-14.39)	4.07 (2.23-5.93)	6.85 (2.29-11.42)	4.85 (4.33-5.38)	0.44 (0.01-0.99)	1.11 (0.63-1.58)	0.49 (0.1-1.096)	0.21 (0.01-0.43)	0.08 (0.06-0.09)
European	-	12.85 (10.31-15.41)	-	2.69 (2.18-3.21)	-	-	0.35 (0.28-0.98)	-	-

UCVA: Uncorrected visual acuity, PVA: Presenting visual acuity, BCVA: Best corrected visual acuity, CI: Confidence interval, WHO: World Health Organization

Table 3 presents the prevalence of the causes of VI and blindness. In the selected articles, refractive errors, with a prevalence of 77.20% (95% CI: 73.40%–81.00%), were the most common cause of VI. Amblyopia, retinal disorders, congenital cataract, and corneal opacities were other causes of visual impairment, and cataract, glaucoma, and refractive errors were the most common causes of blindness.

Table 3.

The proportion (%) of causes of visual impairment and blindness in the reviewed articles

1st author	Causes visual impairment (%)					Causes of blindness (%)
	Refractive errors	Amblyopia	Congenital cataract	Corneal opacity	Retinal disorder	Glaucoma	Refractive errors	Cataract	Glaucoma
Al Faran et al., 199346	67.9	1.3	20.6	3.8	0.6	1.0	5.3	52.6	5.3
Ajaiyeoba et al., 200744	66.6	-	-	-	-	-	-	-	-
Adhikari et al., 201443	1.9						-	-	-
Alrasheed et al., 201647	57.0	5.6	3.7	0.9	13.1	-	-	-	-
Beiram, 197148	-	-	-	-	-	-	-	3.2	10.9
Darge et al., 201752	77.3	4.5	4.5	-	-	-	-	-	-
Demissie and Solomon, 201153	-	-	-	-	-	-	17.0	33.0	11.0
Dorairaj et al., 200854	-	-	-	-	-	-	-	28.7	-
Farber, 200356	-	-	-	-	-	-	4.1	2.7
Fotouhi et al., 200729	87.3	13.2	0.5	0.8	0.5	-	-	-	-
Gilbert et al., 200826	-	30.0	3.3	6.6	36.6	-	-	-	-
Goh et al., 200532	89.5	2.9	0.2	0.1	0.2	-	-	-	-
He et al., 200728	96.8	1.4	0.24	0.24	0.36	-	-	-	-
He et al., 200433	95.6	2.8	0.1	0.1	0.2	-	-	-	-
He et al., 201417	89.5	10.1	0.1	-	-	-	-	-	-
Ibrahim et al., 201318	89.0	5.5	-	-	4.1	-	-	-	-
Jamali et al., 200961	62.1	37.9	-	-	-	-	-	-	-
Kedir and Girma 201464	54.0	5.4	2.7	8.1	10.8	-	-	-	-
Kingo and Ndawi, 200967	31.2	-	-	-	-	-	-	-	-
Kumah et al., 201319	88.8	4.5	1.1	2.3	2.2	-	-	-	-
Lu et al., 200924	80.3	4.2	4.2	-	-	-	-	-	-
Maul et al., 200039	62.1	9.0	0.72	0.48	2.5	-	-	-	-
Murthy et al., 200236	80.9	6.4	0.37	1.3	5.1	-	-	-	-
Naidoo et al., 200335	66.4	9.4	2.3	4.7	10.9	-	-	-	-
Paudel et al., 201415	92.7	2.2	0.7	-	0.4	-	-	-	-
Pi et al., 201272	86.1	9.7	0.42	-	-	-	-	-	-
Pokharel et al., 200038	55.1	12.3	2.9	4.4	5.1	-	-	-	-
Robaei et al., 200531	69.0	22.5	-	-	2.8	-	-	-	-
Salomão et al., 200978	76.8	11.4	-	-	5.9	-	-	-	-
Sapkota et al., 200825	93.3	1.77	0.10	-	1.25	-	-	-	-
Sewunet et al., 201479	87.7	-	-	-	-	-	-	-	-
Srivastava and Verma, 197882	-	-	-	-	-	-	-	32.0	25.0
Taylor et al., 201085	56.0	-	-	-	-	-	33.0	-	-
Thulasiraj et al., 200334	-	-	-	-	-	-	-	-	10.2
Wu et al., 201388	96.6	2.2	-	0.05	-	-	-	-	-
Yamamah et al., 201514	-	0.4	0.4	-	0.4	-	-	-	-
Zainal et al., 200290	48.3	-	35.9	2.5	2.8	-	-	-	-

DISCUSSION

Our study is the first to generate a more accurate estimate of the global prevalence of VI in children using credible population-based studies. We also presented the prevalence of VI and blindness based on different definitions. Studies in the under-20 year's old groups and especially studies in the under-15 year's old groups were the most important reason for choosing 20 years-old as a cut-off. Our results indicated that the lowest prevalence of BCVA VI was 0.057% in the study by Flanagan et al.58 in Ireland and the highest prevalence was 7.29% in a study by Maul et al. in Chile.39 The lowest and highest prevalence of VI based on PVA was, respectively, 19.29% in the study by Adhikari et al.43 and 0.1% in the study by Congdon et al.27 Despite the lower prevalence of VI in children compared to adults (3.82% versus 35.8%10), the number of years lost due to disabilities caused by vision impairment in children imposes a large burden on societies, especially in less developed countries. In a systematic review, Köberlein et al.92 reported that the direct costs of VI included hospitalization, utilization of medical services, purchase of medical products, and the recurrence of VI. They showed that in several population-based studies using representative populations in the United States, the annual cost was 12,175-14,029 dollars for a patient with moderate VI, and 14,882–24,180 dollars for a blind person.92 The high cost of treatment and follow-up on the one hand, and the mental burden, the educational failure, and in general, the reduced quality of life for children on the other hand justify the importance of determining estimates of the trend of the prevalence of VI and its causes in children.

In addition to imposing costs, the burden of disease is an important issue. In a retrospective study, examining data from 195 countries between 1995 and 2017, the disability-adjusted life year (DALY) number of refractive errors in school children was higher than preschool and teenagers.93

Determining the prevalence of VI and its most important causes are necessary to apply policies and strategies to prevent and eliminate the preventable causes of VI. Our findings showed that refractive errors were the most common cause of VI in most articles reviewed in this meta-analysis, such that 29 articles described refractive errors as the cause or one of the causes of VI with rates ranging between 48.3% in the study by Zainal et al.90 and 96.8% in the study by He et al.28 Failure to use the protocol recommended for Refractive Error Study in Children (the RESC Protocol which suggests the use of cycloplegic refraction) in some studies has led to different estimates of the prevalence of refractive errors. In the RESC study, the following definition is defined to determine the refractive error Cycloplegic Refraction: In eyes with successful cycloplegia, refraction is performed with either an autorefractor or retinoscope. Autorefraction is carried out according to the manufacturer instruction manual, including daily calibration. Retinoscopy is carried out using a streak retinoscope in a semi-dark room, with the examiner at a distance of 0.75 meters and a +1.50 diopter lens in the trial frame. Therefore, not using the same definition in studies has led to different estimates in the reports. In studies on similar age groups in geographic regions close to each other, different definitions of refractive errors have been used, and the prevalence of refractive errors, as a cause of VI, is significantly different.51 Another cause of the difference in the prevalence of refractive errors can be the difference between the studied age groups in the reviewed articles. In studies conducted in age groups over 7 years, the refractive errors as a cause of VI is higher than in studies where the average age of the participants is <7 years. In studies such as those by Sapkota et al.25 and Paudel et al.15 where the average age is 10 years and older, over 90% of VI is due to refractive errors. The age-related increase in the prevalence of myopia is one of the major causes of the high prevalence of refractive errors in studies that sampled older age groups. The meta-analysis by Rudnicka et al.94 in the Middle East Region suggested a significant age-related increase in the prevalence of myopia, such that rates changed from 3.5% in the 5-year age group to more than 47% in the 18-year age group. In a trend analysis from 1990 to 2017, the prevalence of children aged 1–14 years with refractive disorders was 1.8% (95% uncertainty interval [UI]: 1.5–2.1). In school children, teenagers, and preschool children, the prevalence was 2.1% (95% uncertainty interval [UI]: 1.5–2.8), 2% (95% UI: 1.4–2.7) and 1.6% (95% UI: 1.2–2), respectively.93 Another cause of difference in the results of these studies can be race and ethnic differences, and thus, genetic and lifestyle differences. In the meta-analysis by Rudnicka et al.,94 the prevalence of myopia in the East Asian Region was more than 80% while it was <5.5% in black African children of the same age group. This racial difference has also been observed with other causes of VI such as amblyopia.

According to our findings, amblyopia is the second leading cause of VI after refractive errors in the reviewed papers. In countries where such screening programs have been in effect for a longer time, the prevalence of amblyopia, as one of the most important preventable causes of VI has been reported. In the absence of apparent strabismus, amblyopia is usually not easily identifiable in children, thus, only properly designed and implemented screening programs by trained people will be effective for the timely diagnosis of amblyopia. Otherwise, childhood amblyopia will continue until they reach adulthood and will lead to a decline in the quality of life in adolescence and older age. Findings by Høeg et al.95 show that the prevalence of amblyopia in the Danish 20 to 29-year old population, who had been screened by the national screening program for children and treated in childhood was 0%, and in cohorts over 50 years of age, the rate was more than 1.5%. This significant difference clearly shows the impact of the implementation and expansion of screening program in recent years compared to previous years.

Based on our findings, the overall global prevalence of blindness in the under 20-year population was 0.17%. The definition by the WHO is based on BCVA <0.05 (20/400). The prevalence of blindness in the studies using this criterion was estimated at 4.5%. These definitions in different countries have always led to various estimates of blindness. For example, blindness is defined as a visual acuity of ≤0.02 (20/1000) in Germany and ≤0.05 in Israel.56,96 Rosenberg and Klie97 have shown that changing the definition of blindness from ≤0.1 to <0.1 can reduce the diagnosis of blindness by up 32%. Establishing national registries for the blind is very important and effective in determining the prevalence and causes of blindness. Unfortunately, few countries have established reliable registries so far, and in other countries, relevant information, such as the prevalence and causes of blindness, is generated from surveys or studies in schools for the blind, and due to methodological errors in these studies, the results are interpreted with caution. This lack of consistency in the definition and diagnosis of blindness and the lack of registries has led to overestimation or underestimation of global blindness. Despite these differences, we determined the prevalence of blindness based on different diagnostic criteria by referring to the most reliable survey articles and excluding studies performed at schools for the blind. Studies have shown that despite the reduction in age-standardized prevalence of blindness and VI over the past 20 years, based on corrected vision, cataract is still the most important cause of blindness in the world, such that in 2015, Khairallah et al.98 reported that more than 33% of the world's blindness was due to cataract between 1990 and the end of 2010. In our study, cataract was the most common cause of blindness and the third most common cause of VI in the reviewed studies. Due to lack of information such as nonreporting standard error or CI, meta-analysis of other causes was not possible for the authors. In 2002, Zainal et al.90 reported the highest prevalence of cataract (3.92%) in children younger than 19 years of age. In determining the cause of blindness and comparing it among different populations, the study of the economic status of the countries and the availability of public health services plays an important role. In countries where access to cataract surgery due to lack of equipment, lack of experienced specialists, and financial inability of people for access to surgery, cataract plays a major role in blindness. In light of this discussion, to reduce preventable blindness, it is necessary to conduct nationwide surveys to determine the existence and availability of surgical facilities and to give priority to raising public awareness for the utilization of healthcare services.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

APPENDIX

Appendix 1: Search methods

The search strategy was created using the following phrase

(Vision impairment or Low Vision or Visual Disorders or Visual Disorder or Visual Impairments or Vision Disability or Visual disability or Vision Disabilities or Day Blindness or Reduced Vision or Subnormal Vision or Diminished Vision or vision impaired or Visual defect or Visual loss or Visually impaired or Visually impaired persons or blindness or Acquired Blindness or Complete Blindness) and (prevalence or epidemiology or cross-sectional stud* or observational stud* or survey). Three international databases including Scopus, Web of Science, and PubMed were searched for publications indexed up to January 2018. To access more articles and to ensure the correctness of the search strategy in the databases, we also reviewed the reference lists of the selected articles as well as Google Scholar.