Thyroid Disease Is Associated with Higher Age-Related Macular Degeneration Risk: Results from a Meta-Analysis of Epidemiologic Studies

Age-related macular degeneration (AMD) is a prevalent disease leading to blindness and involves the loss of a person’s central field of vision. Globally, the prevalence of AMD, which is the leading cause of irreversible loss of vision among elderly individuals, is expected to increase with population aging, ranking third as a cause of blindness after cataract and glaucoma [1]. AMD is a complex multifactorial disease accounting for 8.7% of cases of blindness worldwide, and its prevalence is increasing and spreading [2]. The global population is expected to grow to approximately 9.5 billion by 2050, and it was estimated that there were nearly 2.4 billion people ≥60 years in 2050 [3]. The prevalence of AMD was 2.4% in African Americans, 4.2% in Hispanics, 5.4% in Whites, and 4.6% in Chinese individuals aged 45–85 years, and the estimated total prevalence of any AMD was 4.2% [4]. The pathophysiological mechanisms leading to AMD are not fully understood, but genetic tendencies and environmental factors, including the use of tobacco and nutritional factors, are believed to play an important role [5]. Recently, thyroid hormone (TH) signaling has been implicated in the pathogenesis of AMD [6].

Human retinal pigment epithelial cells express TH receptors and seem to be a direct target for TH [7]. Population studies have revealed that high serum thyrotropin levels are associated with retinal arteriolar narrowing [8]. Animal studies have demonstrated that moderate levels of pre- and postnatal TH insufficiency lead to alterations in visual function in adult rats [9]. These findings highlight the important role of TH in regulating various visual functions.

Ma et al. [10] reported that suppressing TH preserves cone photoreceptors in mice with retinal degeneration, while administrating active TH deteriorates cones, which are known to be important factors for the development and progression of AMD. A Rotterdam study found that higher serum-free thyroxine values were associated with an increased risk of AMD [11]. Furthermore, some studies have demonstrated an association between the use of thyroid medication and the risk of AMD [12].

Previous studies have shown some evidence of a relationship between TH and AMD in the general population, although the findings have been inconsistent. In light of the above, the objective of the present meta-analysis is to evaluate the available studies related to thyroid disease and medication and AMD. Our meta-analysis analyzes relevant information published in the literature to more accurately assess the association between thyroid disease and medication and AMD.

We performed a literature search regarding the association of thyroid disease and medication with AMD; epidemiologic studies published from inception to March 2020 were examined. An electronic search was performed using EMBASE, PubMed, and Google Scholar. The articles were searched using the following keywords in English. The following medical subject headings were used: (“thyroid function” OR “hyperthyroidism” OR “hypothyroidism” OR “thyroid disease”) AND (“age-related macular degeneration” OR “AMD” OR “ARMD” OR “age-related macular disease” OR “age-related maculopathy” OR “ARM”). Searches were restricted to human studies, and all references of the selected articles were searched manually to identify relevant original studies and citations.

The inclusion criteria for the literature were as follows: (1) had a cross-sectional, case-control, or cohort design; (2) evaluated the association between thyroid medication or thyroid function and AMD risk; (3) performed with human subjects; and (4) had adequate data for extraction or calculation, which may be presented as the odds ratio (OR) or relative risk (RR) and 95% confidence interval (CI). The exclusion criteria were as follows: (1) duplicate articles and (2) articles with incomplete data.

The systematic review was conducted by 2 investigators who independently reviewed the titles and abstracts of all studies identified, and disagreements were resolved by consensus. We used a unified data form. The following information was recorded for each study: year, region, study design, sample size, population characteristics, main results, and matched or adjusted variables in the study design or data analysis.

The pooled RRs with corresponding 95% CIs were considered for the effect size analysis in all analyses. When studies had multiple adjustment models, we extracted those reflecting the greatest extent of adjustment for potential confounders. Some studies separated risk estimates according to the different types of AMD (early AMD, advanced AMD, and dry AMD), and any report of AMD effect was included. The heterogeneity test was performed by using Cochran’s Q and I² statistics [13]. Statistical significance for heterogeneity was considered if p value was <0.05 or I² was >50%, and the summary estimate on the basis of the random-effects model is presented; otherwise, a fixed-effects model was performed. Publication bias was estimated using funnel plot graphics and Begg’s and Egger’s tests.

According to the inclusion and exclusion criteria, a total of 13 studies were included in our systematic literature search. A flow diagram that shows how we located relevant studies is presented in Figure 1. Table 1 summarizes the baseline characteristics of the included studies. Seven [11, 14-19] studies involved thyroid disease, and 7 [12, 18, 20-24] studies involved thyroid medication. It is worth noting that data for both thyroid disease and medication were extracted from 1 article [18].

Seven studies with a total of 61,993 patients reported the relationship between thyroid disease and AMD. The presence of thyroid disease was associated with a 25% increase in the risk of AMD in humans (summary RR: 1.25; 95% CI: 1.02, 1.54) (Fig. 2). Statistically significant heterogeneity was observed in the study results (I² = 80.1%, p < 0.001).

Seven studies with a total of 118,420 patients reported the relationship between thyroid medication and AMD. The random-effects model was used for meta-analysis, and the results showed a pooled RR (95% CI) for the association between thyroid medication and AMD of 1.26 (0.92–1.72), which was heterogeneous (I² = 69.0%, p = 0.004), as shown by the forest plots (Fig. 3).

Egger’s test was used to evaluate publication bias. There was no indication of a publication bias for the thyroid disease (p = 0.889) or thyroid medication (p = 0.226) results, as shown by the funnel plots (Fig. 4, 5).

In sensitivity analyses, we recalculated the combined results by excluding 1 study per iteration. Overall, the results of the sensitivity analysis were similar to those of the original analysis, and the RR estimate did not change drastically in the meta-analysis of thyroid disease, thyroid medication, and AMD (Fig. 6, 7).

TH signaling has been reported to regulate cell proliferation, differentiation, and metabolic homeostasis and to be involved in different visual functions, such as the function of specific cone subtypes in human retinal organoids [25]. Thyroid dysfunction and subclinical thyroid dysfunction are common in the general population, and the mean prevalence of total thyroid dysfunction in Europe was found to be 3.82% [26].

However, the mechanisms that link thyroid disease and AMD are currently unclear. It can be speculated that elevated TH levels can accelerate oxidative metabolism by inducing mitochondrial enzymes [27]. The retina is particularly susceptible to oxidative stress, and oxidative damage is thought to be one of the multiple components of the complex molecular mechanism of AMD [28]. In addition, thyroid dysfunction is related to vascular function, lipids, and atherosclerosis [29], which are known predisposing factors for the development and progression of AMD [30]. Population studies have revealed that rs943080 in the VEGF-A gene is associated with AMD and is significant in thyroid-stimulating hormone genome-wide association studies [11]. A possible explanation may be that susceptibility genes have been found in inflammation and oxidative stress pathways. Our findings suggested that thyroxine medication use also appears to play a role in influencing the risk of incident AMD. A population-based study in the region of Augsburg reported a significantly higher risk for other ophthalmic diseases in males treated with TH [24].

The results of the present meta-analysis were based on 7 articles and indicated that the prevalence of thyroid disease was associated with AMD (pooled RR [95% CI]: 1.25 [1.02–1.54]). In addition, there was no statistical association between thyroid medication and AMD risk. This is the first study investigating the putative association between thyroid function and AMD in the general population.

Notwithstanding these numbers, it is difficult to evaluate what the clinical consequences of this heterogeneity may be for future settings. Therefore, we estimate the prediction interval for the RR [31]. The resulting prediction interval for RR of thyroid disease and the risk of AMD, ranging from 0.253 to 2.25, can be interpreted as the 95% range of true RR to be expected in similar studies. The prediction interval contains values above 1 which means that the presence of thyroid disease was associated with an increase in the risk of AMD. But, it also contains values below 1 which means that the presence of thyroid disease may decrease the risk of AMD. Consequently, the result in a new study may be even the exact opposite to the estimate of the meta-analysis. On the other hand, we estimate the prediction interval for the RR of thyroid medication and the risk of AMD, ranging from 0.19 to 2.33. The prediction interval is similar to the 95% range of RR.

There are several limitations in our study. First, a thyroxine drug effect on AMD requires a long period of exposure, and this study had insufficient information in this regard. Second, some studies adjusted for potential confounders. There was no distinction between adjusted and nonadjusted when combining effect values. Last, too few studies have reported AMD type, and consequently, it was not possible to conduct subgroup analyses according to AMD type. Further studies are needed to validate our findings and to better elucidate the influence of TH and medication use on the development of AMD.

Our results indicate that there was an increased incidence of AMD in patients with thyroid disease. Thyroid disease prevention strategies may have a significant effect on the prevention of AMD and warrant further evaluation.

The PRISMA guidelines are given in online suppl. material 1 (for all online suppl. material, see www.karger.com/doi/10.1159/000515273).

All analyses were based on previous published studies, and thus no ethical approval and informed consent are required. Nonetheless, the study adhered fully to the Declaration of Helsinki.

The authors have no conflicts of interest to declare.

The study was supported by Anhui Provincial Department of Education Project (2019jyxm1009).

L.T. conceived the study. M.Z. and Q.Z. performed and checked the available information from eligible articles in this meta-analysis. Z.X. analyzed the data and wrote the main manuscript text. T.X. reviewed and revised the manuscript. All authors read and approved the final manuscript.