INTRODUCTION

Visual impairment (VI) is becoming a major public health concern with the aging of the population. It is the third most prevalent physical impairment among older adults⁽¹⁾. Visual impairment affects one-third of elderly subjects, with the prevalence of mild or severe VI ranging from 4% to 12%^(2,3). Ultimately, untreated eye disorders can contribute to avoidance of social situations, resulting in social isolation or leading to substantial physical decline, which may ultimately require a transition into an assisted living arrangement⁽⁴⁾. A number of studies suggest that disability may contribute to depression⁽⁵⁾; depression is the most common psychiatric disorder and carries a high burden in terms of treatment costs, effect on families and carers, as well as loss of workplace productivity⁽⁶⁾. The World Health Organization (WHO) currently ranks it as the third most prevalent moderate and severe disabling condition globally. It may become a chronic disorder with ongoing disability, particularly if inadequately treated, contributing to the global burden of disease⁽⁷⁾. Results from a number of population-based^(8-12)and hospital-based studies^(13,14) indicate that VI is associated with higher rates of depression.

This review aimed to verify whether there is an association between VI and depression in the elderly.

METHODS

The PRISMA guidelines⁽¹⁵⁾were followed for this review, and the sections below are set out according to these guidelines.

Articles that reported depression, as analyzed by the Geriatric Depression Score (GDS) or another specific scale, in adults aged 65 years and older with VI (visual acuity worse or equal to 20/70 or 0.55 in logMAR) or mild VI (visual acuity worse than 20/40 or 0.3 in logMAR) were included⁽¹⁶⁾. Cross-sectional, cohort studies, or case-control studies were also included. Any setting was permitted (primary care, secondary care, or general population), although the search was limited to articles in English or Portuguese. Review studies, unpublished articles, abstracts, theses, dissertations, and book chapters were not included. Articles regarding dual sensory loss (hearing and VI) were excluded, as were studies about a single ocular disease and those not reporting about VI as the primary objective.

The primary outcomes associated with the aim of this review were depression, most frequently evaluated using GDS⁽¹⁷⁾ or the Center for Epidemiologic Studies-Depression scale (CES-D)⁽¹⁸⁾ and visual acuity. Secondary outcomes analyzed included sociodemographic variables (age, gender, marital status, ethnicity, years of education, living alone or not, social activities), visual function (VF questionnaires), visual field, disability [analyzed by the Community Disability Score (CD-30) or Activities of Daily Living (ADL)], chronic disease, mortality, anxiety [measured using the State-Trait Anxiety Index (STAI-6) or General Health Questionnaire Score (GHQ-28)], physical activity, quality of life [measured using Short-Form Health Surveys (SF)], number of antidepressive medications, previous history of depression, duration of institutionalization if present, and duration of VI.

To gather as many studies as possible on the issue, the strategy for data collection aimed to identify articles that presented the definition of VI and depression in the elderly. The following electronic databases [LILACS, SciELO, MEDLINE (via PUBMED), and Cochrane Central Register of Controlled Trials (CENTRAL)] were searched, from inception to August 2014. Two independent reviewers conducted the search using the same keywords. The search strategy used the following terms: (visual impairment OR blindness) and (elderly) and (depression). Cited references were also evaluated.

Abstracts thought to be relevant were retrieved for closer inspection. If meeting the inclusion criteria, the full text of the article was located. The reference lists of all full-text articles identified during the electronic search were scanned for further articles that may be relevant. In addition, we scanned all related references of the full-text articles meeting the inclusion criteria. Data were independently extracted by two reviewers to ensure accuracy using a standard data table extraction that includes details about the studies such as name, author, year, periodical and database, country, number of participants, study design, variables, limitations, and main results. All included studies were subjected to methodological critical appraisal using the Cochrane Collaboration of risk of bias: the Newcastle-Ottawa Scale, adapted version⁽¹⁹⁾. The original Newcastle-Ottawa Scale⁽²⁰⁾ is used to analyze quality in observational studies. It has eight questions that access three dimensions: selection, comparability, and results (in cohort studies) or exposition (case-control studies). For each question, there are several options, and each option that is associated with the best quality is scored with a star; the more stars a study has, the better is its quality.

In this review, the questions were adapted to investigate exposition and outcome (depression), and the risk of bias was classified as low, unknown, and high risks; each star represents a low risk of bias. Two independent reviewers assessed the quality of studies.

For dichotomous outcomes, we intended to use the odds ratio (OR) and 95% confidence interval (CI) calculated using a random-effect model (REM). When the effect is absent, the risk difference (RD) and 95% CI is calculated using REM. For continuous outcomes, the mean and standard deviation are used to generate the mean difference (MD) and 95% CI using REM. BioEstat version 5.3 was used to conduct meta-analyses. Statistical heterogeneity (I²) was assessed using heterogeneity tests, i.e., the standard chi-square test (p value <0.10 or <10%) and the I² test (I² >50% was considered statistically significant).

RESULTS

Study Selection

Of the 641 search results, 42 full-text articles were retrieved. The reference lists of each of these articles were scanned, and further 20 articles were identified as being possibly relevant. Of the total studies identified, 10 were considered relevant and have been evaluated (Figure 1).

Figure 1 Flow diagram summarizing the process for selecting original articles for review 

DISCUSSION

Depression is a major public health concern among the elderly. It causes suffering, family disruption, disability, worsening of the outcome of many medical illnesses, and increased mortality⁽²⁹⁾. The prevalence of depression among the elderly varies. Almost 3% of the general elderly population suffers from major depression; 8%-16% of the elderly have clinically significant depressive symptoms⁽³⁰⁾.

Rates of VI in the elderly increase with age and are associated with functional loss, mortality as well as psychological distress, sense of loneliness, and depression. Given that depression is treatable and some ocular diseases that cause visual loss are reversible, early identification and treatment of persons most at a risk could have an important impact in the well-being of the elderly⁽²⁶⁾.

The studies included in this review were observational, with half of them being cross-sectional, and these study designs do not show causality. Therefore, it is difficult to confirm a temporal relationship between depression and VI. The limitations of this review were the inadequate definition of outcomes and lack of responses by some authors regarding missing data.

The sociodemographic variables were not associated with depression in most articles. In one study, men were more depressed than women⁽²⁸⁾. In another, women had a higher incidence of depression than men⁽²³⁾. Another study showed that a low level of education is associated with greater symptoms of depression⁽²⁵⁾.

The outcome depression was analyzed using GDS-15 in six studies^{(4,21,24,25,27,28)}; nevertheless, the cut-off for depression varied in these studies, limiting the comparison between heterogeneous groups.

Only one study analyzed a previous history of depression or the use of antidepressant medicine, which is, by itself, a risk factor for depression⁽²⁸⁾.

Depression was related to VI in all studies (prevalence ranged from 8.8% to 45.2%).

Visual acuity was accessed using different methods (Snellen, Glasgow, and ETDRS), and only one study⁽²¹⁾ evaluated the Visual Function Score (VF-14), and this variable was associated with depression. The same study also assessed high- and low-contrast visual acuity, and both were related to more depression. Only one article assessed near and far visual acuity⁽²⁶⁾ and found that only far visual acuity was associated with depression.

Most articles used the best-corrected visual acuity; nevertheless, just one article⁽²³⁾ related that refraction was done when visual acuity was low, considering that refractive errors are reversible VI⁽³¹⁾.

The criteria to determine VI varied between the studies. One study used ICD-9⁽²⁸⁾, while another used ICD-10⁽²⁵⁾. Most articles did not establish standard criteria to diagnose VI. Consequently, there was a high heterogeneity between the VI groups, ranging between worse than 20/40^(23,24) and worse than 20/70⁽²⁵⁾.

The relationship between severity of VI and depression was just reported by two studies^(4,25). One study found that the severity of visual loss does not matter and that VI was a risk factor for depression only in participants without a previous history of depression, and this study found that a 2-year history of depression is the most important risk indicator of depression⁽²⁸⁾. The other study found that there is a positive association of more severe visual loss with depression⁽²⁵⁾.

One study used a questionnaire with subjective questions about the vision of the participant and what the participant was and was not able to do because of the visual loss. This type of questionnaire could be a risk because there is no objective method to measure the real visual acuity in these cases⁽²⁷⁾.

Only two authors evaluated baseline and follow-up visual acuities, and they found that the follow-up visual loss was associated with depression^(26,27). The two studies that analyzed the time of visual loss and depression found different results^(4,25).

Anxiety was analyzed by two authors, and one found a relationship between anxiety and depression and VI; however, the other did not described such an association^(21,24).

Other variables (quality of life, limitation of diary activities, physical activities, and disability) that were previously considered as very important in the association between VI and depression were correlated with depression; nevertheless, they were analyzed in only few studies⁽²¹⁾. When advanced age is coupled with low vision, a person’s level of activity is narrowed further than with either age or VI alone. Furthermore, although older people can normally adjust well to environmental challenges, when the demands on their body are increased, such as when vision is disrupted, physical limitations may become problematic and affect quality of life⁽²¹⁾.

The limitation of daily activities was analyzed in two studies, measured using different methods, and it was associated with depression in both^(21,27).

In one study⁽²⁴⁾, the author reported that a high limitation in daily activities is associated with an eight times greater VI and 12 times more depressed people.

One study⁽²⁶⁾ reported that the time of institutionalization, when it occurs, is not related with depression. In addition, one study related that living alone is not associated with an increase in depression rates⁽²¹⁾.

Disability is already well known in the literature as an important risk factor for VI and depression⁽²⁾, and it is a major public health issue for the elderly. Age is associated with decreased physical competence and increased prevalence of chronic illness⁽²¹⁾. VI is an important, but not life threatening, chronic illness and cause of disability⁽³²⁾ and has been associated with depression in three studies^(12,22,27). Different indices were used to assess disability. Low functional ability was associated with depression in one study, as measured using the Barthel index⁽²⁵⁾.

Mortality and chronic diseases were described in one and two studies, respectively^(23,28); both associated this factor with depression, and these factors are also described in the literature as being important to depression and VI^(33,34). VI causes impaired psychosocial functioning with impaired ADL, loss of independence, reduced social interaction, and depression, and it is known to be related to mortality. In addition, patients may be exposed to shared risk factors for mortality and VI, such as aging, diabetes, vascular pathology, high blood pressure, or smoking⁽²⁸⁾.

Chronic diseases were assessed by the CDC score in one study and were associated with depression and VI; another study reported that stroke and cardiovascular diseases are related to depression^(23,28), and there is a strong association in the literature with these risk factors⁽¹⁸⁾.

The causes of VI were described in one study⁽²²⁾, which demonstrated that diseases that cause central visual loss are more associated with depression than those that present with peripheral visual loss. Another study just reported the prevalence of the diseases⁽²⁷⁾ without specifying whether any of them is associated with depression. None of the articles classified ocular disease as being reversible or permanent, which could have been important in determining whether a person with an irreversible disease had more or less depression than one with a reversible cause.

The implication with regard to clinical practice is that because there is no established link between standardized VI measurements and depression, physicians would have to rely on knowledge and clinical experience to decide when to investigate and treat a depressed visually impaired patient.

This review shows that new research is necessary to analyze each important variable because most studies did not include the full range of common risk factors for both VI and depression, making it impossible to establish a linear correlation between them. In addition, it is important to use standard criteria to classify depression and VI, such as the GDS scale with a standard cut-off to depression and VI criteria following just IDC or WHO.

CONCLUSIONS

After a literature research, we could not definitely establish an association between VI and depression in the elderly. The reasons include a lack of standardized measures of VI and depression to allow comparability of the studies and the potential bias created by several other variables.