Article Text


Epidemiology of blindness in children
  1. Ameenat Lola Solebo1,2,3,4,
  2. Lucinda Teoh1,
  3. Jugnoo Rahi1,4,2,3
  1. 1 Lifecourse Epidemiology and Biostatistics Section, Population, Policy and Practice Programme, Institute of Child Health, University College London, London, UK
  2. 2 Great Ormond Street Hospital/Institute of Child Heath, NIHR Biomedical Research Centre, London, UK
  3. 3 Visual function and integrative epidemiology, Moorfields Eye Hospital and Institute of Ophthalmology NIHR Biomedical Research Centre, London, UK
  4. 4 Ulverscroft Vision Research Group, London, UK
  1. Correspondence to Ms Ameenat Lola Solebo, Lifecourse Epidemiology and Biostatistics Section, Population, Policy and Practice Programme, Institute of Child Health University College London, 30 Guilford Street, London, WC1N 1EH, UK; a.solebo{at}


An estimated 14 million of the world’s children are blind. A blind child is more likely to live in socioeconomic deprivation, to be more frequently hospitalised during childhood and to die in childhood than a child not living with blindness. This update of a previous review on childhood visual impairment focuses on emerging therapies for children with severe visual disability (severe visual impairment and blindness or SVI/BL).

For children in higher income countries, cerebral visual impairment and optic nerve anomalies remain the most common causes of SVI/BL, while retinopathy of prematurity (ROP) and cataract are now the most common avoidable causes. The constellation of causes of childhood blindness in lower income settings is shifting from infective and nutritional corneal opacities and congenital anomalies to more resemble the patterns seen in higher income settings. Improvements in maternal and neonatal health and investment in and maintenance of national ophthalmic care infrastructure are the key to reducing the burden of avoidable blindness. New therapeutic targets are emerging for childhood visual disorders, although the safety and efficacy of novel therapies for diseases such as ROP or retinal dystrophies are not yet clear. Population-based epidemiological research, particularly on cerebral visual impairment and optic nerve hypoplasia, is needed in order to improve understanding of risk factors and to inform and support the development of novel therapies for disorders currently considered ‘untreatable’.

  • Blindness
  • Vision disorders
  • Epidemiology

Statistics from


An estimated 14 million of the world’s children are blind.1 A blind child is more likely to live in socioeconomic deprivation,1–3 to have delayed or disordered development, to be more frequently hospitalised during childhood and to die in childhood than a child not living with blindness.1 4 5 The differential between the blind and non-blind child is more pronounced in developing nations: while 10% of UK children die in the first year following the diagnosis of blindness, in lower income countries, the equivalent mortality is 60%.2

This article updates our previous review on childhood visual impairment (VI),6 by summarising new evidence on the global epidemiology of, and the emerging therapies for, severe visual impairment and blindness (or SVI/BL; table 1). It is now recognised that in adults even mild VI is associated with lower socioeconomic status and poorer general and mental health status.7 The evidence base regarding children is inconclusive. Moderate VI (table 1) may impact on educational opportunities, with half of the UK’s moderately visually impaired children educated within specialised schools for children with physical or learning deficits.8 However, there is still a paucity of research on the epidemiology and impact of childhood moderate VI. For example, it is well recognised that the majority of blind children will have other non-ophthalmic disorders or impairments,5 but it is unknown whether the same is true of those with moderate VI.

Table 1

WHO International Classification for Disease classification of visual impairment (VI), severe visual impairment (SVI) and blindness (BL)62

Defining blindness

The 1972 WHO taxonomy still forms the basis of the International Classification for Disease definition (table 1) of blindness.9 The recent creation of an additional diagnosis of ‘monocular blindness’ is important as these individuals have a lifelong increased risk of binocular blindness due to visual loss in the seeing eye,10 but the impact on global development for children with monocular blindness is unclear.

All humans are born with vision below adult acuity norms, with the average neonate having acuity worse than 1.0 logMAR. This improves rapidly in the first year of life (figure 1).11 12 There is no child-specific taxonomy for visual disability. As age and cognition may be obstacles to quantification of a child’s acuity level, childhood SVI and BL are often categorised together (SVI/BL).2 5 Children diagnosed in the first year of life, who constitute the majority of blind children in many populations, invariably have clinical signs consistent with very poor vision, such as absence of preferential looking behaviour when presented with high contrast visual stimuli or obvious severe ocular anomalies.5 Although children with SVI/BL often have similar causative disease profiles and similar ocular phenotypes, by definition, this grouping includes both children with vision sufficient to navigate around the world independently (eg, 1.1 logMAR) and those with absolutely no perception of light. The life experiences, cognitive, developmental and educational outcomes for children at the two ends of this spectrum may differ, and until very recently there was a paucity of child-centric measures of experiences and outcomes.13 14

Figure 1

Maturation of vision in the first 2 years of life derived from Salomão and Ventura11 and Mayer et al, 12 copyright Solebo and Rahi.

Global burden of childhood blindness

The major challenge to quantifying burden is the rarity of the disability and the individual causative conditions. Population-based approaches are required to capture a representative picture. Additionally, there are varying definitions of both childhood (WHO: <14, Unicef: <16 and UN Convention of the Child: <18) and blindness.

Much of the literature on the epidemiology of childhood blindness is based on study populations drawn from schools for children with disabilities or children seen within healthcare centres.6 These methodologies are often at risk of underascertainment and bias particularly in lower income settings,15 where there are significant obstacles to accessing education or healthcare.16 There is evidence that research has failed to capture girls, rural communities or children with multisystem impairments.15 16 Children may also fail to present to healthcare services because families do not recognise that there is a problem17 or because access to healthcare for children is limited by their carer’s own blindness.17 There can also be a lack of awareness among healthcare givers, with half of a group of primary care workers in Tanzania unaware of the urgent nature of referrals for congenital cataract or that children with albinism may have VI.18

Key informant (KI) studies, in which trained volunteers with a pre-existing ‘key’ role identify children with disorders in their community, enabling referral to healthcare professionals, have been validated as a low-cost alternative to other population-based approaches. KI methods enable researchers to capture a more representative study population but are still likely to underestimate the true burden.19

Using the available estimates of childhood blindness, derived through robust population-based approaches, the prevalence of blindness in individuals aged under 16 years (the definition used most consistently within the research) has been estimated at 12–15 per 10 000 children in very poor regions and 3–4/10 000 in affluent areas (figure 2).1 As the birth rate remains higher within lower income countries, these countries have a disproportionately higher absolute number of blind children.1

Figure 2

Global prevalence of childhood blindness presented by economic region. Derived from Rahi and Gilbert.1 Prevalence per 10 000. Numbers of children affected in millions.

Trends in the global causes of childhood blindness

For children in higher income countries, cerebral VI (CVI) and optic nerve anomalies remain the most common causes of SVI/BL, while retinopathy of prematurity (ROP), cataract, glaucoma and non-accidental injury are now the most common avoidable causes.3 20 Recent work from the UK has suggested an increasing certification of children with SVI/BL.21 This suggests either a true increase in new cases or an increasing awareness of the benefit of certification. Certification leads to registration, which brings with it increased support for the child and family (although it is not a prerequisite for access to early educational and developmental support for UK children with visual disabilities). However, this voluntary register has been found to be incomplete.3 Much of the data on the epidemiology of childhood blindness in an industrialised setting is derived from the 2001 British Childhood Visual Impairment Study.5 A follow-up study, which aims to investigate the epidemiology of the full spectrum of childhood VI, is currently underway.22

Over the last two decades, with the establishment of national programmes of vitamin A supplementation, vaccination and sanitation improvements, the constellation of causes of childhood blindness in lower income settings has shifted from infective and nutritional corneal opacities and congenital anomalies to more resemble the pattern of causes seen in higher income settings.16 23 24 In countries that have relatively recently moved up from the lower to the middle economic strata, there has been improved survival following premature or complicated birth, with an attendant increase in visual morbidity, due to ROP and CVI.25 26 Of 231 000 children (aged under 16 years) examined as part of a major recent Indian rural population based study, 8 per 10 000 had vision worse than 3/60 (95% CI 40 to 110/10 000).27 Almost half of the blind children had retinal disorders, the most common being ROP. Cataract (28%) and globe anomaly (11%) were the next most common blinding disorders.27 Among blind children in regions of Nigeria, 30% were blind due to an event in the perinatal period.25 In Turkey, perinatal birth injury-related CVI is now the most common cause of childhood blindness with significant decreases in blindness secondary to ROP and cataract.28

While maternal (vertical) and ‘horizontal’ transmission of potentially blinding diseases such as measles and rubella has fallen, other infectious diseases may come to the fore. For example, a third of Brazilian children with suspected Zika-associated microcephaly have ocular abnormalities, the most common being pigmented change or retinochoroidal atrophy.29 There have also been reports of ocular changes in exposed infants with normal head circumference.30 However, CVI associated with severe microcephaly is likely to be sufficient cause of poor visual function for most affected children, who are also likely to have severe global brain dysfunction

The review will now summarise key developments in the epidemiology and management of the most important causes of childhood SVI/BL (table 2), that is, those that are responsible for the highest proportion of affected children globally and those that carry the highest burden of avoidable blindness.1 3 6 15 21 25 27 28 31–33

Table 2

The most important causes of childhood blindness1 2 9 62 (which may coexist)

Retinopathy of prematurity

ROP develops when the vasoconstrictive response to hyperoxia, that is, the immature retina of the eyes of premature children, is followed by a vasoproliferative phase that is driven by the surge in endothelial growth factors on the return to normal oxygenation. In industrialised settings, CVI is a more important cause of VI for a preterm child, but globally, ROP remains the major threat to vision for preterm infants. Approximately 170 000 preterm babies worldwide developed some degree of ROP in 2010, and 54 000 required treatment for potentially blinding severe disease, but only an estimated 42% of these babies received this treatment.23 Globally, an estimated third of surviving children with ROP requiring treatment (20 000; 95% CI 15 500 to 27 200) are severely visually impaired or blind secondary to ROP.23

Developments in neonatal care in regions such as Western Europe and North America have led to a reduction in the proportion of moderately preterm children (32–28 weeks) developing ROP.24 34 However, in less developed health settings, older and heavier preterm babies are still at significant risk of developing ROP.35 36 Of Iranian children with ROP requiring treatment, 8% would have not been screened if American or UK guidelines had been in place.37 This finding, consistent with others from low-income/middle-income countries, demonstrates the power of epidemiology in determining setting-specific policy and practice.24

Primary prevention of blindness due to ROP requires identification of the risk factors underlying disease development. Prematurity and low birth weight are the most important determinants of disease and may be addressed through maternal healthcare. Genetic and environmental factors are likely to play a role in the degree and duration of hyperoxia necessary to trigger the process, the resultant surge in vascular growth factors and the severity of disease that develops. Balancing oxygen supplementation is key: lower supplementation (85%–89% vs 91%–95%) reduces the risk of sight threatening ROP (risk ratio (RR) of 0.72, 95% CI 0.51 to 1.00) but increases the risk of mortality (RR 1.17, 95% CI 1.03 to 1.32).38 The role of nutrition, haemoglobin transfusion or erythropoietin (EPO) administration is unclear, though meta-analysis of RCTS of early nutritional supplementation indicates a protective effect (RR 0.22, 95% CI 0.09 to 0.55 and RR 0.67, 95% CI 0.46 to 0.97, respectively).36 38

Secondary prevention strategies involve early identification and early and appropriate treatment of children with ROP to prevent blindness. The required infrastructure can be a challenge in higher income settings, where approximately 55 infants are examined for every infant treated.34 Over 8 years, the coordination of care over 36 rural centres in India involved more than 75 000 imaging sessions on more than 23 000 preterm infants.39 One in every 15 examined infants required urgent treatment. Telemedicine techniques may allow more babies to be examined,35 40 41 but suitably trained ophthalmologists are still required to confirm diagnosis and deliver treatment. Reliable prediction of those at risk will be key to delivering a sustainable service: the weight, insulin-like growth factor (IGF)-1, neonatal, ROP (WINROP) study used serum IGF levels and postnatal weight gain to successfully predict all children who required ROP treatment.42 The WINROP algorithm is undergoing validation across different countries and settings to determine its utility as a universal screening tool.

Retinal laser ablation therapy is challenging, time consuming and implicitly destructive but remains the gold standard intervention to prevent central sight loss in children with severe ROP. One in 12 babies undergo disease progression despite laser treatment.34 Antivascular endothelial growth factor (VEGF) agents have recently emerged as a therapeutic option. Bevacizumab (Avastin) is now used as a first-line treatment by up to a quarter of ophthalmologists following early studies suggesting superiority over laser in more central disease.40 However, the long-term neurodevelopmental and cardiovascular impact of the associated suppression of systemic VEGF levels, which can last for up to 8 weeks after intravitreous bevacizumab injection, is unclear. A recent retrospective analysis of a cohort of very premature children showed that, adjusted for maternal education, systemic status and gestational age, children who had undergone intravitreal treatment were, by 6–7 years of age, more likely to have severe neurodevelopmental impairment (RR 3.1, 95% CI 1.2 to 8.4).43


Cataract related to prenatal rubella infection is still an issue globally, for example accounting for 20% of childhood cataract in the Philippines.44 However, for the majority of affected children with bilateral cataract (and therefore at risk of blindness), aetiology is unknown, and prevention of blindness is focused on the prompt detection and treatment of visually significant lens opacity before deprivation amblyopia becomes intractable. The Chinese Childhood Cataract Program, established with the support of the V2020 programme, resulted in earlier diagnosis of cases of congenital/infantile cataract, and an apparent increase in childhood cataract prevalence, as a result of improved case detection in remote regions.45

However, there are still many obstacles to prompt treatment for affected children. In several countries, patients need to supplement health costs, putting treatment beyond the means of many families.46 In many settings, proximity to a hospital is an independent predictor of better visual outcome following childhood cataract surgery.47

Corneal opacity

Corneal opacity secondary to vitamin A deficiency, infection or toxicity from traditional remedies remains the most common cause of childhood SVI/BL in sub-Saharan Africa and areas of extreme deprivation,1 despite recent V2020 programmes on nutritional supplementation, measles and rubella vaccination and health education. The only current treatment for established corneal scarring is corneal transplantation. Childhood corneal transplants have a high failure rate due to rejection, new scar formation or infection.48 They require frequent reoperation, and even when the transplant successfully retains its clarity, the complex refractive aberrations can result in intractable severe amblyopia.48 Stem cell therapeutic approaches to corneal transplantation will reduce the incidence of graft rejection but will not overcome the other challenges of paediatric corneal transplantation.

Cerebral visual impairment

CVI (‘cortical’ and ‘central’ visual impairment are terms previously used by some as an alternative to cerebral) is VI due to a heterogenous group of disorders affecting the optic radiations, visual cortex or associated visual areas. This encompasses a spectrum of visual problems from blindness to children with normal acuity but disabling visual processing defects (such as object confusion or non-recognition). Although CVI can coexist with ophthalmic abnormalities, it can be a challenge to determine whether a child with normal eyes and apparently severely reduced visual acuity has CVI unless non-visual causes for poor visual response, such as severe global brain problems, are excluded. For two-thirds of children who have vision worse than 1.0 logMAR due to CVI, VI is part of a global developmental sequelae to hypoxic ischaemic encephalopathy (HIE).5 49 As HIE is strongly associated with birth complication, primary prevention of CVI blindness requires improvement of maternal and infant perinatal health. As with many other early childhood disorders, the incidence of HIE is a marker for the socioeconomic development of a region, being much lower in a high income setting (6 per 1000 live births)50 than that seen in lower income settings (26 per 1000 live births).50 It is postulated that the first 48 hours are the ‘golden window’ for interventions to prevent further neuronal and white matter injury and that a multitarget approach is necessary to reverse the multiphasic ischaemic, apoptotic, inflammatory and angiogeneic pathways underlying HIE.51 The advent of hypothermia (head cooling or whole body cooling) as a therapy for HIE within the ‘golden window’ has resulted in modest improvements in neurodevelopmental outcomes.51 There are several currently underway trials of adjuvant therapies hoped to further improve outcome, including noble gases (NCT 00934700 and NCT 01545271), melatonin (NCT01862250) and erythropoietin derivatives (NCT 01913340).

Other causes of CVI include central nervous system malformations, neoplasia or infection and metabolic neurodegenerative disease.52 Genome studies have also revealed aetiological insights: a whole exome study of 25 children with CVI and cognitive impairment revealed that 16 children had a related underlying genetic abnormality: five had a recognised genetic cause and 11 had mutations within candidate genes coding for neurometabolic functions or brain/optic nerve development.53

Demonstrable improvements in visual function in some children with CVI have resulted in suggestions that various ‘visual stimulation’ therapies may be of benefit, but the population who improve may be a separate subgroup in whom CVI is a manifestation of a delayed or interrupted rather than aborted trajectory of ‘normal’ visual development.54 Further research into pathogenesis, or the identification of therapeutic targets, is hampered by the absence of a clinically meaningful taxonomy with which to classify the different CVI phenotypes.

Optic nerve anomalies

Anterior visual pathway disorders are responsible for almost a quarter of childhood SVI/BL in some higher income settings, and optic nerve hypoplasia (ONH), is the the most common single cause of SVI/BL in industrialised nations.3 20 55 56 In most cases, the cause is unknown, but ONH is independently associated with younger maternal age and nulliparity.55–57 It may also be a marker of poorer maternal health,55 56 with a UK study finding case clusters within areas of socioeconomic deprivation.58 ONH is a clinical diagnosis based on the appearance of the optic nerve, and the absence of a standardised clinical phenotype for the classification of hypoplasia limits epidemiological research. There is evidence of the relatively frequent coexistence of ONH and CVI, but the aetiological or clinical significance of this is unclear.52 There is a paucity of normative data on optic nerve appearance in early childhood or optic nerve volume on neuroimaging during childhood. Hand-held optical coherence tomography devices, which are non-contact diagnostic tools able to produce biomicroscopical images of the paediatric eye, are an emerging technology that may be able to aid the classification of paediatric optic nerve disease.59

Inherited retinal disorders

Photoreceptor dystrophies are the most common inherited retinal disorders among children with SVI/BL.5 60 These constitute the global photoreceptor dystrophy of Leber’s amarosis (LCA), dystrophies affecting rod photoreceptors more than cones (the retinitis pigmentosas) and the cone dystrophies.5 The RPE65 gene, mutations of which cause LCA type 2 and retinitis pigmentosa, has been a target for gene therapies. Following intraretinal injection of adenoviral delivered copies of functioning RPE65, children with LCA2 initially had improved visual function. This improvement was not maintained in follow-up studies due to degeneration of treated retina.61 Further human trials of genetic therapeutics are underway. Next-generation sequencing technologies have resulted in a deeper understanding of the genetic bases of this group of disorders, but as genetic heterogeneity is a challenge to gene therapy, further work on mutation independent approaches is necessary.


Childhood VI and blindness remain an important public health issue, and alongside local or disease specific successes, there has been an emergence, or re-emergence, of other causes of early onset VI, particularly ROP (in middle-income settings) and CVI (within higher income settings). Improvements in maternal and neonatal health and investment in and maintenance of national ophthalmic care infrastructure are key to reducing the burden of avoidable blindness. Therapeutic targets are emerging for childhood visual disorders, although novel therapies for diseases such as ROP or retinal dystrophies are not without risk and the hypothermic therapies that address CVI are still at an early stage. In order to reduce the burden of childhood blindness attributable to diseases previously considered ‘untreatable’, particularly CVI and ONH, population-based epidemiological studies are needed. Such research will determine natural history and putative risk factors, necessary for the elucidation of the pathogenetic mechanisms that will form the basis of future treatments.


View Abstract


  • Contributors All authors were involved in the synthesis of review findings. ALS drafted the initial manuscript. LT and JR critically reviewed the manuscript. All authors can take responsibility for the integrity of the data.

  • Competing interests None declared.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Atoms
    Martin Ward Platt