Non-genetic risk factors for keratoconus

ABSTRACT Keratoconus is a complex and multifactorial disease and its exact aetiology remains unknown. This current study examined the important environmental risk factors and their association with keratoconus. This study was registered in the PROSPERO International Prospective Register of systematic reviews under registration number CRD42021256792 in 2021. Scopus, Web of Science, PubMed, and Cochrane CENTRAL databases were searched for all relevant articles published from 1 January 1900 to 31 July 2021. National Institutes of Health Quality Assessment Tool was used to assess the methodological quality of the studies. The assessment for statistical heterogeneity was assessed using the Z-statistics on RevMan v5.4. P-value of <0.05 was considered as statistically significant and I2 < 25% as homogenous. Thirty studies were included in this meta-analysis. Pooled odds ratio was calculated with 95% CI. The pooled odds ratio (OR) of eye rubbing, atopy, asthma, and eczema was 3.64 (95% CI, 2.02, 6.57), 1.90 (95% CI, 1.22, 2.94), 1.36 (95% CI, 1.15, 1.61) and 1.90 (95% CI, 1.22, 2.94), respectively. The OR for diabetes was 0.86 (95% CI 0.73, 1.02), and use of sunglasses, contact lens, allergic conjunctivitis, side sleep position and prone sleep position was 0.40 (95% CI, 0.16, 0.99), 1.68 (0.70, 4.00), 2.24 (95% CI, 0.68, 7.36), 3.81 (95% CI, 0.31, 46.23), 12.76 (95% CI, 0.27, 598.58), respectively. Twenty studies were considered to be of high quality, nine to be moderate and one to be low. Environmental risk factors have been identified to play a role in the susceptibility of keratoconus. However, further large-scale longitudinal studies are needed to understand the mechanisms between environmental risk factors and keratoconus.


Introduction
Keratoconus is a complex multifactorial degenerative disorder characterised by progressive focal thinning and corneal ectasia. 1 Patients suffering from the disease initially experience distortion and blurring of vision caused by irregular astigmatism, myopia, and corneal scarring. 2 If left unnoticed and untreated, keratoconus can result in severe vision impairment and eventual blindness. Currently, there exists high variability among patients with regard to the clinical manifestation and disease progression.
The exact aetiology of keratoconus remains to be fully understood, but as with all complex diseases, evidence points to a multifactorial aetiology combining genetic and nongenetic or environmental factors. 3 A hereditary link has been identified in keratoconus, with multiple genetic systemic and ocular disorders linked to keratoconus, including Down syndrome, Leber's congenital amaurosis, retinitis pigmentosa, and Marfan's syndrome. 2,[4][5][6][7] However, no causative gene(s) has yet been found. The development of keratoconus most likely requires environmental influences such as eye rubbing, contact lens wear, ultraviolet light exposure, sleeping position, diabetes, atopy, and allergic diseases. 8,9 Of this variety of environmental factors, some more than others have been extensively investigated and reported, 10,11 but there has not been a comprehensive systematic review that evaluates all environmental factors. By identifying most of the environmental risk factors of keratoconus, this would help reduce the incidence of this condition by redirecting efforts towards control of these factors, while also reducing its economic burden.
This systematic review and meta-analysis aims to investigate environmental risk factors and their possible associations with keratoconus.

Methods
This study was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) statement. The protocol of this meta-analysis was published online at the International Prospective Register of Systematic Reviews (PROSPERO) under registration number (CRD42021256792).

Inclusion criteria
Primary studies were considered eligible for inclusion in this review if they met the following criteria: • Full original articles • Published from 1 January 1900 to 31 July 2021 • English language only • Studies involving human beings only • Randomised controlled trials and non-randomised observational studies (cohort and case-control studies) assessing environmental risk factors for the development and progression of keratoconus.
There was no limit on the population group in terms of age, sex, ethnicity or co-morbidities.

Exclusion criteria
• Studies assessing environmental factors without reporting a direct association between environmental factors and keratoconus • Cross-sectional studies using multivariate analysis to explore potential risk factors. Such studies are more hypothesis generating rather than hypothesis testing • Review articles, case reports, surveys, PowerPoint presentations, abstract-only studies, and studies without full-text available were excluded

Study selection
Two authors (MS and QYF) independently screened the titles of publications for the inclusion and exclusion criteria and all potential studies were noted. The titles and abstracts were read to further filter the included studies. The complete texts of the studies were then obtained and read in full to verify final inclusion. Any disagreement was resolved by reaching a consensus through discussion.

Data extraction
Information was extracted from the studies included in this work. Data extracted included the title, authors, date, country of origin, study design, demographics, sample size, disease definition, assessment of environmental factors, covariates adjusted, results including risk measurements, and p-values. Any disagreement was resolved by reaching a consensus through discussion other authors.

Outcome measures
The outcome of the study was risk estimates of eye rubbing, contact lens wear, atopy, asthma, atopic dermatitis, allergic conjunctival diseases, blepharitis, sleeping position, diabetes, and ultraviolet light exposure with respect to keratoconus in terms of odds ratio (OR) calculated at 95% confidence interval (CI).

Quality assessment of the methodology of included studies
The methodological quality of the 30 studies was assessed using the National Institutes of Health Quality Assessment Tool. Depending on the study design, the corresponding tool was used; for example, the tool for Observational Cohort and Cross-Sectional Studies was used for cohort studies. The tool for cohort studies comprised 14 questions, while that of case-control studies included 12 questions, where each could be answered yes; no; cannot determine (CD), not applicable (NA), or not reported (NR). The tools do not specify a certain cut-off score for low-or high-quality methodology. Therefore, if the answer was 'yes' to >67% of questions, the study was considered to be high quality, 33% to 67% to be moderate quality and <33% as low quality.

Statistical analysis
The meta-analysis was performed using review manager ('RevMan', version 5.4, Cochrane collaboration, Oxford, UK). Dichotomous data outcomes were analysed using a Mantel-Haenszel random-effects model and results presented as an OR with 95% CI. Statistical difference between the groups was considered to be present if the pooled 95% CI did not include 1 for the respective OR. All p-values were two-tailed and considered statistically significant if less than 0.05. For each factor analysed, a forest plot showing the respective OR or standardised mean differences with their corresponding 95% CI for each study and for the pooled data were generated. The test of overall effect was assessed using the Z-statistics on RevMan v5.4 with statistical significance set at p < 0.05. Heterogeneity (inconsistency) between studies was evaluated using the Cochrane Q (Chi 2 test) and I 2 statistics in RevMan v5.4. The statistical significance for heterogeneity using the Chi 2 test was set as p < 0.10. Estimates of degree of heterogeneity using I 2 were made by setting 25%, 50%, or 75% as limits for low, moderate or high heterogeneity, respectively. A randomeffects model with weighting of the studies was used when there was heterogeneity between studies (p < 0.10) and I 2 values of over 50%. According to the Cochrane Handbook for Systematic Reviews, the heterogeneity was selected at p < 0.10 rather than conventional level of p < 0.05, as the tests are often not sensitive and wrong exclusion may occur. Therefore, a higher cut-off provides greater evidence for clinical heterogeneity. Egger's test was carried out to evaluate the possible presence of significant publication bias.

Results
A total of 1665 potentially eligible records were extracted in the initial data retrieval process. During screening, 1127 records were eliminated due to duplication, and 441 were eliminated based on the study title and abstract. Of the 97 full text articles reviewed, 67 were excluded for not meeting the inclusion criteria. Thirty studies were finally included for quantitative assessment in meta-analysis (Table 1). Egger's test for a regression intercept gave a P-value of 1.000, indicating no evidence of publication bias. The process used to search and identify studies is illustrated in Figure 1.   OR of allergic conjunctivitis was found to be 2.24 (95% CI, 0.68, 7.36). The OR of developing keratoconus with side sleep position was 3.81 (95% CI, 0.31, 46.23), while that with prone sleep position was 12.76 (95% CI, 0.27, 598.58) (Figures 2-4).

Quality assessment
The quality of the studies included was assessed using the National Institutes of Health Quality Assessment Tool. This review involves 8 cohort studies and 22 case-control studies,   Questions are tailored to the study and they reveal the risk of potential for selection, information and measurement bias, or confounding variables. Tables 2-4 portray the responses of studies to each question. Table 2 includes responses to eight questions that are common to tools of case-control studies and cohort studies. Table 3 has six questions that are specific to cohort studies and Table 4 outlines those specific to casecontrol studies. Overall, 20 studies were considered to be of high quality, 9 to be moderate and 1 to be low.
Studies investigating multiple risks simultaneously or observing keratoconus progression in order to identify risk factors, rather than choosing a certain risk to explore, had lower quality scores. Therefore, future studies should aim to first identify the main environmental risk factor that will be in question before data collection, keeping in mind the key potential confounding variables. Although there is no strict recommendation on the sample size, a justification on the number should be provided. For example, the case-control study by Kuo et al. 12 recruited participants after calculating the minimum sample size needed to detect the protective risk of diabetes for a 2-tailed alpha of 0.05 and a power of 0.80.
Cohort studies tended to have higher quality assessment scores than case-control studies. The discrepancy between the two types of studies generally occurred due to a lack of information or standardisation of controls. If conducting a case-control study for further research, scrutiny is required for clear and comprehensive information on both patient and control groups, ensuring concurrent controls are used and that controls are selected or recruited from the same or similar population as the keratoconus patients. An example of this criteria being fulfilled is the study by Lin et al., 13 which matched the date of medical visit for each keratoconus patient to a non-keratoconus patient (control). Such matching ensures that the date of keratoconus diagnosis is concurrent with that of medical enrolment for controls. Were key potential confounding variables measured and adjusted statistically for their impact on the relationship between exposure(s) and outcome(s)?

Discussion
Various environmental risk factors such as eye rubbing, atopy, asthma, eczema, diabetes, allergic conjunctivitis, blepharitis, sleeping position, and use of sunglasses and contact lens were associated with keratoconus. Findings from this study stipulate for further large-scale longitudinal research in these specific environment risk factors to evaluate the geneenvironment hypothesis of keratoconus. 10,11

Eye rubbing
Eye rubbing has been identified as a well-known significant risk factor for keratoconus. Although, the exact mechanism by which eye rubbing initiates and/or induced keratoconus progression is unknown, it is hypothesised that sheer stress from vigorous and prolonged eye rubbing causes mechanical fatigue of the cornea. 14,15 Pooled risk estimates from all the included studies showed the OR as 3.64 (95% CI, 2.02, 6.57). However, there are inconsistencies in the definitions and criteria for eye rubbing. Studies have grouped the different aspects of techniques, frequencies, durations, and forces under the broad term of eye rubbing. 14,16 It is, therefore, necessary to characterise the various types and techniques of eye rubbing as studies have shown that only repetitive, forceful, and prolonged rubbing of the eye may alter corneal biomechanics to a significant extent to thereby cause corneal trauma and influence the evolution of keratoconus. 11,14,16,17 It is also important to specify the different types of eye rubbing and the rubbing force produced by the techniques as they may be of significance when collecting data from study participants. Moran et al. 17 identified the different patterns of rubbing techniques, including rubbing with knuckles, rubbing with the base of thumbs, rubbing with fingertips, beneath the eyes, or inside the eye. Among the different patterns of rubbing techniques, knuckle rubbing generates the greatest force. 14,16 Furthermore, Moran et al. 17 reported that knuckle rubbing contributes most significantly (OR = 8.29; 95% CI = 3.92-18.26) to the method of eye rubbing, followed by rubbing with fingertips (OR = 5.34; 95% CI = 2.44-12.21). 17 Therefore, the literature has been consistent in reporting that keratoconus subjects are more forceful when rubbing their eyes compared to non-keratoconus subjects. 14,16 This places greater importance on force rather than technique in the development of keratoconus which should be further explored as an area of future research.
Frequency of eye rubbing is another important element in the consideration of keratoconus development. This is because frequent rubbing may prevent the propensity for the ocular surface to recover between periods of rubbing episodes. 16 This is consistent with a study that reported a higher prevalence of keratoconus in the Middle East with an increased frequency of eye rubbing. 14 Thus, there is a need to establish a reliable measure for frequency to gain better insight into how frequency plays a role in the aetiology and progression of keratoconus.

Contact lens wear
Contact lens wear has been reported to be associated with the development of keratoconus. It is thought that wearing a contact lens produces ocular changes, including central corneal thinning, decreased keratocyte density and squamous metaplasia. 18 However, Macsai et al. 19 reported that although contact lens wear inflicts trauma to the cornea, it is not severe enough for the development of keratoconus in non-susceptible patients.
On the other hand, susceptible patients who wear contact lens along with other risk factors that involve ocular trauma such as Down's syndrome, Leber's congenital amaurosis, atopy and significant history of eye rubbing, may cause keratoconus. 19 Additionally, contact lens wear can lead to bulging of the central cornea and flatter keratometry readings compared to noncontact lens wear subjects. Corneal hypoxia has also been reported due to contact lens wear, especially hard polymethylmethacrylate lenses, but the association between the two is unclear and thus warrants further exploration as an area of research. In a similar vein, Barr et al. 18 found that contact lens wear increases the risk of incident corneal scarring by almost 3 times over no contact lens wear.
Gasset et al. 20 investigated the role of hard contact lens wear as an environmental risk factor in the development of keratoconus and concluded that this association is casual one rather causal. Pooled generated in this work reveals a risk association of 1.68 (0.70, 4.00). Therefore, further studies with more definite control for confounders and well-designed cohorts or interventions are warranted. Contact lens wearers should be followed up closely in the detection and monitoring of keratoconus signs.

Atopy
Atopy refers to a genetic predisposition to developing allergic diseases such as allergic rhinitis, asthma, and atopic dermatitis (eczema) due to heightened immune responses to common allergens. Whilst the overall pooled OR was 1.90 (95% CI, 1.22, 2.94), the association between keratoconus and atopy is conflicting in the literature.
Several studies have reported a positive association, 2 including a prospective observational study of 200 consecutive patients presenting with keratoconus by Weed et al. 21 By contrast, a population-based retrospective study by Lee et al. 22 concluded that atopy has no significant association with a positive keratoconus diagnosis. The inconsistencies in the literature may have arisen due to the different definitions of atopy used in these studies. The study by Weed et al. 21 evaluated all the clinical manifestations of this hypersensitivity reaction, including asthma, eczema, and hay fever, whereas Lee et al. 22 only assessed eczema but neither asthma nor hay fever. This then influences the reason as to why atopy may be significantly or insignificantly associated with keratoconus, with a positive association explained by a history of severe rubbing due to atopy, and an insignificant association elucidated by the lack of severe eye rubbing to develop keratoconus.
There is a need to establish a reliable and homogenous definition of 'atopy' to gain better insight into whether or not atopy plays a role in the development and progression of keratoconus.

Allergic diseases
Allergic diseases, which are clinical manifestations of atopy, have been suggested to play an essential role in the aetiology of keratoconus. Their association with keratoconus is well described in the literature and depending on the geographic and demographic origin of the studies, studies have reported various associations between keratoconus and allergic diseases. The allergic diseases that will be explored for their relationship with keratoconus include asthma, eczema, allergic conjunctivitis, keratoconjunctivitis, and blepharitis.

Asthma
There is variability in the association between asthma and keratoconus in different studies, and the exact causal association between the two is still unclear. A positive association has been reported by many authors, including a populationbased matched case-control study by Lin et al. 13 (OR = 1.18; 95% CI = 1.07-1.30), a prospective age-and sex-matched case-control study by Naderan et al. 23 (OR = 4.448; 95% CI = 2.594-7.630) and a prospective observational study by Weed et al. 21 In contrast, other studies which have not found a significant association between asthma and keratoconus. IGordon-Shaag et al. 2 found that asthma was insignificantly different between keratoconus and control groups (crude OR = 2.6, p = 0.16). Similarly, Lee et al. 22 reported that the proportion of patients with asthma was lower in a keratoconus group (P = 0.028), a conclusion consistent with other studies from the Middle East. The present study found the overall OR as 1.36 (95% CI, 1.15, 1.61).

Eczema
The association between eczema and keratoconus is also conflicting: a positive association has been reported by many authors. Overall associated risk as pooled from the included studies was 1.90 (95% CI, 1.22, 2.94), which was in line with the study by Naderan et al. 23 who investigated various risk factors of keratoconus through a prospective case-control study. These authors found a higher prevalence of eczema in keratoconus patients compared to control subjects (OR = 2.976; 95% CI = 1.676-5.282).
Weed et al. 21 evaluated 200 consecutive keratoconus patients and found the prevalence of eczema to be significantly higher than a control group. This is explained by an increase in the frequency of eye rubbing in eczema sufferers within a keratoconic population. Other studies have failed to find a significant association between eczema and keratoconus. Naderan et al. 24 evaluated 922 keratoconus patients and controls and revealed a low prevalence (2.4%) of eczema in keratoconus patients compared to control subjects. Similarly, Lee et al. 22 reported eczema as having no significant association with keratoconus in a Korean population, and Gordon-Shaag et al. 2 did not find any significant differences between keratoconus and control groups regarding eczema (crude OR = 6.2, p = 0.1).

Allergic conjunctival disease
Allergic conjunctival disease is a broad group of allergic diseases involving inflammation of the conjunctiva. It includes seasonal allergic conjunctivitis, atopic keratoconjunctivitis, and vernal keratoconjunctivitis, which vary in their degree of severity and threat to eyesight. However, one thing in common is the positive association between increased keratoconus development risk in patients with allergic conjunctival diseases. The overall risk estimate was found to be 2.24 (95% CI, 0.68, 7.36).
Naderan et al. 2,23 evaluated the effect of allergic diseases on keratoconus and reported an increased prevalence of vernal keratoconjunctivitis and allergic conjunctivitis in keratoconic patients. They also found that patients with allergic conjunctival diseases had more severe keratoconus in comparison with keratoconus patients without allergic diseases. 23 Lee et al. 22 reported 37% increased odds of developing keratoconus (OR 1.37; 95% CI = 1.12-1.67) in subjects with allergic conjunctivitis. Wang et al. 25 found alterations in the corneal tomography, biomechanics and epithelial thickness in allergic conjunctivitis patients which are consistent with the corneal changes seen in patients with early keratoconus. This demonstrates the possibility of allergic conjunctivitis as a risk factor for the progression to keratoconus. Weng et al. 26 also concluded that there is an increased risk of developing keratoconus in atopic keratoconjunctivitis patients. Therefore, patients who experience allergic conjunctival disease should be closely monitored and advised to routinely undergo corneal tomographic and biomechanical measurements for early keratoconus screening.

Blepharitis
The association between blepharitis and keratoconus has been studied by Mostovoy et al. 27 in a prospective, comparative, and observational study. The results reported a higher prevalence of signs and symptoms of blepharitis in keratoconus subjects compared to the control group. Even though the aetiology of keratoconus is not exactly known and most likely multifactorial, blepharitis is thought to play a role in the pathogenesis of keratoconus due to chronic inflammation and persistent eye rubbing that blepharitis causes. 27 Further investigation of the role of blepharitis-related inflammatory mediators and eye rubbing in the pathogenesis and progression of keratoconus is warranted to establish a full understanding of the association between blepharitis and keratoconus. This will also guide the treatment of underlying blepharitis-related inflammation to prevent the development of frank keratoconus and its deterioration. 27

Sleeping position
Sleeping position has recently been identified as an environmental risk factor for initiating and/or inducing the progression of keratoconus. Mazharian et al. 14 were the first to report a significant association between keratoconus and inappropriate sleeping position, i.e. prone and side sleep position, with direct contact of the eye on the pillow ('pillow hugging'). They found a higher prevalence of keratoconus in patients who had incorrect sleeping positions (OR = 14.31; 95% CI = 4.78-42.84).
The pathogenesis of incorrect sleeping position leading to the development and/or progression of keratoconus is thought to be due to a compression of the globe and a generation of increased heat locally in the eye during sleep. 14 Moran et al. 17 also demonstrated significant associations between prone and side sleep positions, and keratoconus. From the present pooled estimate measures, OR of developing keratoconus with side sleep position was 3.81 (95% CI, 0.31, 46.23), while that with prone sleep position was 12.76 (95% CI, 0.27, 598.58). A potential explanation for this is the generation of substantial mechanical stress and local irritation when there is direct contact between eye and pillow. This is supported by the finding of supine sleeping position being a protective factor against keratoconus. 17 As such, further investigations are needed to evaluate the long-term stability of keratoconus after adopting a correct sleeping position.

Diabetes
Recent literature has focused on the potential connection between diabetes mellitus and keratoconus. The two most recent meta-analyses in 2021 and 2022 conducted by Xing-Xuan et al. 28 and Akowuah et al. 29 respectively, uniformly concluded no significant association between diabetes and keratoconus but found that diabetes might have a protective effect for keratoconus. An overall estimate of OR was found to be 0.86 (95% CI 0.73, 1.02). The meta-analyses included studies conducted by Lin et al., 13 Woodward et al., 30 Naderan et al., 24 Kuo et al., 12 and Seiler et al., 31 almost all of which were retrospective in nature. Therefore, more prospective studies assessing the association between diabetes mellitus and keratoconus are needed. Furthermore, the protective effect of diabetes mellitus on keratoconus may be explained by the induction of crosslinks by glycosylation of corneal fibres due to elevated glucose, thus strengthening the cornea and reducing the risk of developing ectasia and keratoconus. 28 Since there is consistent evidence in the clinical literature regarding the association of diabetes mellitus and keratoconus, there is a need to establish viable therapies aimed at raising glucose levels locally in the cornea to prevent and treat keratoconus. 28

UV exposure
The association between sun exposure and keratoconus is equivocal. Gordon-Shaag et al. 2 found that wearing sunglasses had a protective factor against keratoconus (OR = 0.40; 95% CI = 0.19-0.84), whereas those without sunglasses had an increased risk of keratoconus. Overall risk association of keratoconus and sunglass usage amongst the included studies was estimated to be 0.40 (95% CI, 0.16 -0.99). This supports the theory that UV exposure plays a role in the pathogenesis of keratoconus as UV is a source of reactive oxygen species that can cause oxidative stress. 2 Gordon-Shaag et al. 5 reported a higher prevalence of keratoconus in countries with greater sun exposure compared to countries with much less sun exposure. However, the different prevalence of keratoconus in geographic regions can be explained by ethnic differences rather than the amount of sun exposure. 5 Therefore, further studies are required to ascertain whether sun exposure plays a role in the pathogenesis of keratoconus.

Conclusion
Environmental risk factors associated with keratoconus examined in this study included eye rubbing, contact lens wear, allergic conjunctival disease, blepharitis, and sleeping position. Further research is indicated to investigate these risk factors relationship and better explain the pathogenesis, as more recent papers are changing the understanding of potential keratoconus pathogenic mechanisms. A greater understanding of the preceding factors for keratoconus may allow for earlier recognition and prevention of those at risk and therefore lower the incidence rates of this pathology that bears significant impacts on quality of life.

Disclosure statement
No potential conflict of interest was reported by the author(s).