Association between visual impairment and cognitive disorders in low and middle income countries: a systematic review

Authors: Gbessemehlan A., Guerchet M., Helmer C., Delcourt C., Houinato D., Preux PM.

Geographical coverage: Asia, Latin America and sub-Saharan Africa

Sector: Burden of disease

Sub-sector: Epidemiology

Equity focus: Older population.

Study population: Older people in LMICs (specific age-cut off not defined).

Review type: Other review

Quantitative synthesis method: Systematic review

Qualitative synthesis method: Not applicable

Background: Visual impairment and cognitive disorders are common among older people in low and middle income countries (LMICs). However, no synthesis of current knowledge exists for LMICs.

Objectives: Synthesise evidence on the association between visual impairment and cognitive disorders among older people

Main findings:

In the qualitative synthesis, eight publications met the eligibility criteria and were included. These articles were conducted in various regions: five in Asia (China, Thailand, Bangladesh, and Singapore), two in Latin America (Brazil, Cuba, Dominican Republic, Venezuela, Mexico, and Peru), and one in sub-Saharan Africa (Tanzania). The participants’ average age ranged from 64.2 to 76.2 years, with a majority being female. It’s noteworthy that only three studies specifically investigated the correlation between visual impairment and cognitive disorders. Among the eight studies, four found a significant association, two suggested a potential link, and two did not report any statistically significant effect. The authors also observed considerable variability in the methods used to assess visual and cognitive impairments.

In general, the studies were of high quality, with NOS scores for the cross-sectional studies ranging from 7/10 to 10/10. The primary factors compromising the quality of these studies were the inadequate assessment of visual impairment (either due to the lack of validated tools or the use of a simple question) and insufficient or missing information on assessment methods. This was particularly evident in studies where the primary objective was not to investigate the relationship between visual and cognitive impairments. Additionally, there was a lack of systematic presentation of rates or summary characteristics of the non-respondent population or samples.

As very few studies explored the association between visual and cognitive impairments among older people in LMICs, further studies are required to improve the knowledge on this relationship. Authors further note that feasibility or pilot studies assessing to what extent visual impairment correction could reduce the risk of dementia and improve its management in the LMIC context should be encouraged, especially in Africa where the evidence is still insufficient.

Methodology:

The review does not explicitly state the inclusion criteria. However, it seems that studies were considered if they discussed a general or specific visual impairment in conjunction with a low- or middle-income country (LMIC). There was no specified age limit for the populations included in the studies.

The authors clarified that studies were excluded if they were not relevant (such as those involving an inappropriate study population) or did not investigate the association of interest. For instance, studies examining the link between ocular diseases like glaucoma, cataract, DMA, diabetic retinopathy and cognitive disorders were excluded. Similarly, studies that solely evaluated the association with other sensory impairments (such as hearing or olfactory) were also not considered.

This systematic review, adhering to PRISMA guidelines, conducted a literature search on the correlation between visual and cognitive impairments in Low-and Middle-Income Countries (LMICs). Databases such as EMBASE, Medline, Global Health, PsycInfo, PsycArticles, and Web of Science were utilized, with specific search terms related to visual impairment, cognitive disorders, and LMICs. Additionally, grey literature and the database of the Institute of Tropical Neuroepidemiology in University of Limoges were searched using “dementia” as a keyword. The search was then narrowed down to records exploring the association with visual impairment in LMICs. The search was completed on 14 February 2019

n this systematic review, two independent investigators used the Rayyan App to screen titles and abstracts. They retrieved full-texts and compared their choices for inclusion or exclusion. In case of any disagreement, they held meetings to reach a consensus on the articles to be included or excluded. For included studies, data was extracted by the first investigator and checked by the second investigator.

To evaluate the methodological quality of included studies, authors used the Newcastle-Ottawa Scale (NOS) was used to evaluate study quality. Overall score ranges from 0 to 10 for cross-sectional and case-control studies, and from 0 to 9 for longitudinal studies. Scores obtained were categorised into four groups: articles with an overall score below 4, from 5 to 6, from 7 to 8 and from 9 to 10 were respectively considered as poor, medium, good and very good quality.

Applicability/external validity:

The authors highlight that given the limited number of studies examining the relationship of interest and the diverse definitions of visual impairment and cognitive outcomes, it is challenging to draw a definitive conclusion about the strength of the association between these disorders in LMICs, and to compare this with the evidence from high-income counrties.

Geographic focus: Study limited to LMICs articles, and the majority of those included are located within Asia.

Summary of quality assessment:

The methodologies employed for identifying, including, and critically evaluating studies were notably solid, with multiple authors participating in each phase and a search strategy that was not limited by language or publication status. However, there was no mention of consulting relevant experts during the search process. Although the analysis approach for the included studies was rigorous, there was no effort to separately analyze studies based on their associated bias risks. Consequently, we hold medium confidence in the review’s findings.