Causal Relationship Between Common Allergic Diseases and Keratoconus: A Two-Sample Mendelian Randomization Study

Abstract

Purpose: To investigate the impact of several allergic diseases on the development of keratoconus (KC) from a genetic perspective.

Methods: Data from multiple genome-wide association studies involving European ancestry individuals were used. Ultimately, 613 allergic conjunctivitis (AC) cases and 474,657 controls, 4387 allergic rhinitis (AR) cases and 471,273 controls, 4859 allergic asthma cases and 135,449 controls, 1169 allergic urticaria cases and 212,464 controls, 22,474 atopic dermatitis cases and 774,187 controls, and 311 KC cases and 209,287 controls were included. Then inverse-variance weighted (IVW) method, potential pleiotropy, heterogeneity, and leave-one-out analyses were assessed to investigate the causal effects of those allergic diseases on KC.

Results: IVW estimates indicated that allergic asthma promoted the development of KC (IVW odds ratio [OR] = 2.18; 95% confidence interval [CI], 1.007–4.707; P = 0.048). However, the IVW estimates show that AC (IVW OR = 1.02; 95% CI, 0.960–1.086; P = 0.513), AR (OR = 0.26; 95% CI, 0.037–1.754; P = 0.165), atopic dermatitis (IVW OR = 1.50; 95% CI, 0.942–2.395; P = 0.088), allergic urticaria (IVW OR = 1.11; 95% CI, 0.822–1.511; P = 0.485) are not significantly causally associated with KC.

Conclusions: Allergic asthma may promote KC by increasing allergy-related ocular inflammation.

Translational Relevance: The possibility of increased KC in patients with allergic asthma should be considered.

Introduction

Keratoconus (KC) is a progressive corneal ectasia characterized by thinning and conical protrusion of the central and paracentral cornea, culminating in irregular astigmatism and significant visual deterioration.^1,2 The prevalence of KC varies markedly across populations, with estimates ranging from 0.2 to 4790 cases per 100,000 individuals, and the highest incidence reported in the 20–30 year age group and among Middle Eastern and Afro-Asian populations.¹ KC is a leading cause of visual impairment in young adults. Despite considerable research efforts, the pathogenesis of KC remains poorly understood, and current therapeutic approaches predominantly focus on mitigating disease progression. The initial clinical signs of KC are subtle, often mimicking refractive errors such as myopia and astigmatism, which can delay diagnosis. Moreover, KC has been linked to several risk factors, including positive family history, frequent eye rubbing, and atopic conditions such as eczema, asthma, and allergic rhinitis (AR).^2–4 Associations with systemic conditions, including Down's syndrome and connective tissue disorders, have also been documented.⁵ Therefore a comprehensive understanding of the etiological factors underlying KC is essential. 

Atopic disorders—including allergic conjunctivitis (AC), AR, asthma, atopic dermatitis, and allergic urticaria—are characterized by aberrant immune responses resulting from a complex interplay of genetic and environmental factors.⁶ Emerging evidence suggests that allergic mechanisms play a significant role in the pathogenesis of KC.⁵ Chronic eye rubbing, commonly observed in patients with AC, is hypothesized to be a critical factor contributing to the development of KC. Studies indicate that patients with KC are at an increased risk for developing allergic AC,^7–9 although there are also studies that report no significant association between KC and AC,¹⁰ thus rendering the relationship between these two conditions a subject of ongoing debate. Furthermore, individuals with AR exhibit a higher incidence of KC than those without.¹¹ Woodward et al.¹² found that patients with asthma have a 31% higher risk of developing KC. However, although epidemiological studies have consistently demonstrated associations between KC and allergic diseases, the underlying mechanisms and causal relationships remain largely unresolved.¹³ Thus further research is required to elucidate the role of allergic diseases in the pathogenesis of KC. 

Mendelian randomization (MR) is a powerful genetic epidemiological tool that leverages genetic variants as instrumental variables to infer causal relationships between exposures and outcomes,¹⁴ reducing the potential for confounding bias inherent in traditional observational studies.¹⁵ Owing to the random segregation of alleles during meiosis, MR offers a robust approach to minimizing confounding factors¹⁶ and enhancing the reliability of causal inferences. 

This study used a two-sample MR analysis to investigate the potential causal relationship between KC and allergic diseases and explore common pathogenic pathways. We anticipate that these findings will advance our understanding of the relationship between these conditions and provide a foundation for the development of early diagnostic and therapeutic strategies for at-risk individuals. 

Methods

Study Design and Data Resource

A concise flowchart provides an overview of the fundamental principles of MR (Fig. 1). To address the issues of reverse causation and unmeasured confounding inherent in observational studies, we conducted an MR analysis to investigate the causal effects of several common allergic diseases on KC. 

Figure 1.

 

The following flowchart illustrates the MR study of common allergic diseases and keratoconus. The study postulates a potential causal link between common allergic diseases and keratoconus, although there is currently no evidence to suggest a connection with confounders. Instrumental variables are relevant with exposure but not with regard to confounders or outcomes. Consequently, IVs can exert an effect on the outcome only through exposure.

A two-sample MR study was conducted to explore the causal association between several allergic diseases and KC. The study utilized data from genome-wide association studies (GWAS) involving individuals of European ancestry. The study included data from 613 cases of AC (ebi-a-GCST90018791) and 474,657 controls, 4387 cases of AR (ebi-a-GCST90018792) and 471,273 controls, and 4859 cases of allergic asthma (finn-b-ALLERG_ASTHMA) and 135,449 controls, 1169 cases of allergic urticaria (finn-b-L12_URTICA_ALLERG) and 212,464 controls, and 22,474 cases of atopic dermatitis (ebi-a-GCST90027161) and 774,187 controls, as well as 311 cases of KC (finn-b-H7_CORNEALDEFORM) and 209,287 controls. The aforementioned studies were approved by the local institutional review board and ethics committee. 

Selection of Instrumental Variables

A series of quality control criteria was employed with the purpose of selecting eligible genetic instrumental variables (IVs). The IVs were extracted according to a significant threshold of P < 5 × 10⁻⁸. Subsequently, the clumping procedure was performed with R² < 0.001 and a window size of >10,000 kb to avoid linkage disequilibrium. Furthermore, single-nucleotide polymorphism (SNP) harmonization was performed to guarantee the accurate orientation of the alleles. 

Exclusion of Weak Instruments

To assess the strength of genetic instruments related to common allergic diseases, we calculated F statistics for all SNPs. IVs with F statistics less than 10 were considered weak instruments and were excluded from MR analysis. 

MR Analysis

The MR analysis was conducted in R (version 4.4.1) using the Two Sample MR (version 0.6.6) packages. In this study, the inverse-variance weighted (IVW) method was used as our primary analysis to estimate the causal effects of exposures on the risk of common allergic diseases and KC. Additionally, we performed MR Egger, weighted median, simple mode, and weighted mode as a complementary method. 

Pleiotropy, Heterogeneity, and Sensitivity Analysis

The potential for horizontal pleiotropy was assessed using the MR Egger regression approach. In cases where pleiotropy was identified (P ≤ 0.05), the MR Pleiotropy Residual Sum and Outlier (MR-PRESSO) test was conducted using the MRPRESSO package in R, and the SNP with the lowest pleiotropy P value was removed. Moreover, both the IVW and MR Egger regression approaches were used to evaluate heterogeneity, which was measured using Cochran's Q statistic. A leave-one-out analysis was performed to verify the robustness of the results. 

Results

The objective of this study was to analyze the causal relationship between the most common allergic diseases, namely, AC, AR, allergic asthma, atopic dermatitis, and allergic urticaria, and the progression of KC. 

Causal Effects of Allergic Asthma on KC

A positive association was observed between allergic asthma and KC, with six SNPs demonstrating a statistically significant correlation (odds ratio [OR] = 2.18; 95% confidence interval [CI], 1.007–4.707; P = 0.048) (Fig. 2). The causal associations are visually summarized in scatter plots (Fig. 3). The leave-one-out analysis demonstrates that the findings from the IVW analysis are robust and not influenced by any single SNP (Fig. 4A). Furthermore, the funnel plot indicates that the results are reliable and free from significant publication bias (Fig. 4B). It is noteworthy that our study did not identify any horizontal pleiotropy (intercept = −0.049, P = 0.882). These findings lend support to the hypothesis that there is a causal relationship between allergic asthma and KC. 

Figure 2.

 

Forest plot illustrating the results of MR for common allergic diseases with a potential causal relationship to KC. The reference line is set at an OR = 1; values to the left indicate that the allergic disease acts as a protective factor for KC, whereas values to the right signify that it is a risk factor.

Figure 3.

 

Scatter plot illustrating the causal association between allergic asthma and the risk of KC. The effects of SNPs are represented by lines corresponding to various statistical methods: inverse-variance weighted test (light blue line), MR-Egger regression (blue line), weighted mode (red line), weighted median (green line), and sample mode (light green line). The slope of each line reflects the estimated MR effect for the respective method.

Figure 4.

 

(A) Leave-one-out plot illustrating the causal effect of allergic asthma on KC. The red line represents estimations derived from the IVW test. The leave-one-out analysis indicates that the removal of each individual SNP, one at a time, does not significantly affect the results of the remaining SNPs. (B) Funnel plot of the causal association between allergic asthma and KC.

Causal Effects of Other Common Allergic Diseases on KC

Moreover, the IVW method, employed as the primary analytical approach, did not identify a causal association between KC and other prevalent allergic diseases, including AC (OR = 1.02; 95% CI, 0.960–1.086; P = 0.513). The results demonstrated no significant association between KC and other common allergic diseases, including atopic dermatitis (OR = 1.50; 95% CI, 0.942–2.395; P = 0.088) and urticaria (OR = 1.11; 95% CI, 0.822–1.511; P = 0.485), as illustrated in Figure 2. Given the limited number of SNPs for allergic rhinitis in the GWAS database, we used the Wald ratio method, which indicated that AR had no significant effect on KC (OR = 0.26; 95% CI, 0.037–1.754; P = 0.165) (Fig. 2). 

Discussion

This study represents a pioneering effort to elucidate the potential causal relationship between multiple allergic diseases and KC through two-sample MR analysis, which bolsters earlier observational findings and posits that allergic asthma may play a distinct and significant role in the pathogenesis of KC. By using two-sample MR analysis, we effectively mitigated the influence of confounding factors, thereby clarifying the genetic association between allergic asthma and KC. Not only does this research deepen our understanding of the pathophysiology of KC, but it also lays a scientific foundation for identifying susceptible populations and investigating targeted interventions in clinical practice. Future investigations may uncover the specific pathological mechanisms underlying allergic asthma, offering novel insights into the prevention and treatment strategies for KC. 

Analysis of Causal Mechanisms Between AC and KC

AC is a common ocular surface disorder encountered in the clinical setting, characterized by symptoms such as itchy eyes, conjunctival congestion and edema, tearing, foreign body sensation, and photophobia. A previous study revealed that individuals with AC had a 37% higher likelihood of developing KC.¹⁷ Additionally, research by Merdler et al.¹⁸ has demonstrated a significant correlation between AC and KC. However, the literature suggests that allergies are merely a correlate rather than a causative factor in the development of KC.¹⁹ Our MR analysis results did not find a definitive causal link between KC and AC. The meta-analysis by Seth et al.¹⁰ indicated that KC had no significant association with allergic eye disease, consistent with our conclusion. 

Our analysis highlights several key considerations regarding the relationship between AC and KC. First, many studies are based on specific populations such as the Korean National Health Insurance Service National Sample Cohort database, which may introduce selection bias and limit generalizability across demographic groups. Second, observational studies are also vulnerable to confounding factors such as eye-rubbing behavior, environmental influences, and genetic predisposition, and diagnostic criteria for AC and KC vary by country and socioeconomic or racial variables can further complicate the analysis.^20–22 Third, although there is no direct causal link between AC and KC, AC may indirectly affect the onset and progression of KC through behaviors (such as eye rubbing) or via inflammatory mediators, which may not be adequately captured by the current MR analysis methods. Numerous studies have demonstrated that eye rubbing is highly prevalent among patients with KC,^23–25 which can cause corneal damage, thinning, reduced stiffness, and curvature changes,²⁶ potentially accelerating KC progression. Given the multifactorial nature of KC development, with AC serving as one of several potential contributors, it remains challenging for MR analysis to identify a definitive causal relationship, emphasizing the need for future research to consider these indirect pathways and confounding factors to better understand this relationship. 

Analysis of Causal Mechanisms Between AR and KC

The current research on the relationship between AR and KC remains controversial. Our MR analysis did not reveal a clear causal relationship between AR and KC. However, some existing studies present findings that differ from our results. For instance, a study involving Israeli adolescents reported a significantly increased odds ratio for KC in patients with AR (P < 0.001).¹⁸ Similarly, a bivariate meta-analysis found that patients with AR were significantly more likely to develop KC.¹¹ AR is a non-infectious chronic inflammatory disease, primarily mediated by IgE, with its core mechanism being IgE-mediated type I hypersensitivity. Allergy-related inflammatory processes have been hypothesized to contribute to the pathogenesis of KC. Notably, elevated serum IgE levels were observed in 52% of patients with KC versus compared to only 7% of controls, suggesting a possible association between AR and KC. 

The association between AR and ocular symptoms has received considerable attention in recent years. AR is also associated with ocular inflammatory and irritative symptoms like itching and dry eyes. Several studies propose that KC may be linked to these ocular inflammatory and irritative symptoms through the nasolacrimal duct, which connects the nasal and ocular regions, facilitating the transfer of allergens and inflammatory mediators.¹¹ In addition, patients with AR frequently rub their eyes due to discomfort, which may further exacerbate the progression of KC. 

However, a meta-analysis by Lo et al.²⁷ found no significant correlation between AR and KC. Similarly, Woodward et al.¹² indicated no statistically significant difference in AR prevalence between KC patients and controls. These findings are consistent with our MR analysis, reinforcing the conclusion that no direct causal relationship exists between AR and KC. It is plausible that previous observational studies may have been confounded by factors such as ocular symptoms and behaviors, including eye rubbing, both of which are recognized risk factors for KC. After controlling for these confounding variables, our MR analysis indicates no direct causal link between AR and KC. Thus, although AR may be associated with KC, this association is likely mediated by other factors rather than being directly causal. Further studies are needed to clarify this complex relationship. 

Analysis of Causal Mechanisms Between Allergic Asthma and KC

In our two-sample MR analysis, we found that allergic asthma promoted the progression of KC. Allergic asthma, an inflammatory disease of the small airways in the lungs, is caused by combined genetic and environmental factors.²⁸ Xu et al.²⁹ revealed that allergic asthma (95% CI, 1.10–1.51; P = 0.0014) were strongly associated with a higher risk of KC. Furthermore, a retrospective longitudinal cohort study by Woodward et al.¹² indicated that individuals with asthma have a 31% increased risk of developing KC and an additional 21% increased risk of progression to severe KC. Another study using a database from Asian countries similarly demonstrated that patients with allergic asthma have a higher likelihood of developing KC (OR = 1.18; 95% CI, 1.07–1.30; P = 0.0007).³⁰ Asthma is not only associated with KC but also increases the severity of the condition, as confirmed by Sahebjada et al.²⁸ These findings suggest a strong association between allergic asthma and KC and reinforce a positive causal relationship between the two conditions. 

There is evidence for a shared genetic contribution between asthma and KC. A GWAS has confirmed that genetic variants in the IL-6 receptor gene are associated with asthma.³¹ The ligand for this receptor, IL-6, is a potent pleiotropic cytokine that regulates cell growth and differentiation and plays a crucial role in the immune response. Studies have indicated that IL-6 levels in the tears of patients with KC are significantly higher than those in healthy controls and that disease severity is significantly correlated with IL-6.^32,33 IL-6 also increases the expression of TNF-α through the mitogen-activated protein kinase pathway. IL-6 and TNF-α, among other inflammatory markers of KC, are considered to contribute to increased corneal damage and the progression of KC.^33–35 Although elevated levels of IL-6 and TNF-α in the tears of patients with KC suggest a potential role in disease pathology, they do not definitively establish IL-6 as a pathogenic factor of KC. Our MR results only indicated a causal relationship between allergic asthma and KC; however, the underlying biological mechanisms require further investigation. 

Analysis of Causal Mechanisms Between Atopic Dermatitis, Allergic Urticaria, and KC

Our analysis revealed the absence of a direct causal link between atopic dermatitis, allergic urticaria, and KC. Consistent with our results, a cross-sectional study by Merdler et al.¹⁸ indicated that allergic urticaria and atopic dermatitis were not significantly associated with KC. Although some studies have suggested a correlation between atopic dermatitis, allergic urticaria, and KC,^27,36 an underlying causal relationship is lacking. Our MR analysis results suggest that atopic dermatitis and allergic urticaria are unlikely to directly contribute to the development of KC but may indirectly influence KC through other pathways. For example, corneal mechanical damage or oxidative stress resulting from eye discomfort or rubbing may indirectly contribute to the development of KC.³⁷ These indirect factors may not have been adequately accounted for in the MR analysis, potentially leading to the lack of a significant causal relationship. 

Despite substantial investigative efforts, the genetics, prevention, risk factors, progression, treatment, and underlying pathophysiology of KC continue to be areas of limited understanding. In this study, we investigated the potential causal relationship between allergic diseases and KC through an MR analysis of two distinct sample sets, uncovering a promotional effect of allergic asthma on the development of KC. Although this investigation offers potentially informative insights into the etiology of KC, it is not without limitations. Notably, the reliance of the study on GWAS data from European populations and the lack of MR analysis of Asian populations are important considerations that may have affected the generalizability of our findings. 

In summary, our study is based solely on GWAS databases, utilizing MR to analyze the causal relationships between various allergic diseases and KC from a genetic perspective. As these associations are influenced by multiple factors in reality, including behavioral (e.g., eye-rubbing) and environmental effects, further research is needed to investigate the underlying disease mechanisms and their associations in a multifactorial context. 

However, our findings suggest that allergic asthma may promote the onset of KC likely through increased ocular inflammation. Conversely, AC, AR, atopic dermatitis, and allergic urticaria did not exhibit any promotion of KC. 

Acknowledgments

Supported by National Natural Science Foundation of China (Grant No. 82301240), Tianjin Health Research Project (Grant No. TJWJ2023MS036), and the Tianjin Key Medical Discipline (Specialty) Construction Project (TJYXZDXK-016A). 

Disclosure: X. Yang, None; X. Yu, None; S. Gao, None; P. Wei, None; G. Han, None