Genetic Polymorphisms Affecting Ranibizumab Response in High Myopia Patients

High myopia is an ophthalmic pathology that affects half of the young adults in the United States and Europe and it is predicted that a third of the world’s population could be nearsighted at the end of this decade. It is characterized by at least 6 diopters or axial length > 26 mm and, choroidal neovascularization (CNV) in 5 to 11% of cases. Ranibizumab is a recombinant humanized monoclonal antibody fragment. It is an anti-vascular endothelial growth factor (anti-VEGF) drug used in the treatment of CNV. Many genetic polymorphisms have been associated with interindividual differences in the response to ranibizumab, but these associations were not yet assessed among patients with high myopia and CNV. We performed a retrospective study assessing the association of genetic polymorphisms with response to ranibizumab in patients with CNV secondary to high myopia (mCNV). We included genetic polymorphisms previously associated with the response to drugs used in CNV patients (bevacizumab, ranibizumab, aflibercept, and photodynamic therapy (PDT)). We also included genetic variants in the VEGFA gene. Based on our results, ARMS2 (rs10490924) and CFH (rs1061170) are associated with response to ranibizumab in high myopia patients; and, included VEGFA genetic polymorphisms are not associated with ranibizumab response in our population but might be related to a higher risk of CNV.


Introduction
Myopia is a pathology in which the image is cast in front of the retina instead of on it. It happens because the eyeball is too elongated, or in a minority of cases because of the increasing refractive index of the crystalline in nuclear cataract or cornea in keratoconus [1].
Myopia noted a rise worldwide over time. It affects about half of young adults in the United States and Europe, and it is even more prevalent in Asia. It has twice the prevalence compared to 50 years ago, and it is forecast that a third of the world's population will be nearsighted at the end of this decade [2].
Pharmaceutics 2021, 13, x FOR PEER REVIEW 3 of 19 genetic variant was associated with decreased response to bevacizumab [31]. Also, many genetic polymorphisms in the VEGFA gene were associated with interindividual differences in the responses to anti-VEGFA drugs or PDT among AMD patients [27,29,32]. Among CNV and/or polypoidal choroidal vasculopathy patients, usually considered a subtype of CNV, also there are SNPs associated with variable response to ranibizumab, bevacizumab or PDT as HTRA1-625 A/G (rs11200638), F13A1 (rs5985) CFH I62V (rs800292), CFH Y402H (rs1061170) and, especially, ARMS2 A69S (rs10490924) [33]. As we can see, no studies have assessed the association of genetic variants with interindividual differences in the responses to ranibizumab among high myopia patients despite ranibizumab being commonly used to treat this pathology.
This study aims to evaluate the association of VEGFA variants with interindividual differences in the responses to ranibizumab. It also aims to assess the association of genetic polymorphisms previously associated with the response to treatments used in patients with CNV (anti-VEGF drugs and PDT), excluding drug-SNP interactions in non-ophthalmic indications (cancer, hypertension). Furthermore, as a secondary objective, it is intended to characterize these genetic polymorphisms in our population and evaluate their possible association as genetic markers of CNV.

Materials and Methods
A retrospective study including patients with high myopia and CNV treated with ranibizumab between 2014 and 2019 in San Cecilio University Hospital (Granada, Spain). We studied the association of genetic polymorphisms with differences in response to treatment from baseline (BL) up to 1 and 6 months of follow-up.
The Research Ethics Committee of Granada approved the study (0085-N14; 26 May 2014). We obtained written informed consent from all participants and we followed the principles of the Declaration of Helsinki.

Inclusion/Exclusion Criteria
The inclusion criteria were: 1. Diagnosed high myopia and CNV, 2. Treatment with ranibizumab by prescription from ophthalmologists at our hospital, 3. Treated at our hospital, 4. At least six months follow-up with available medical records. As we can see, no studies have assessed the association of genetic variants with interindividual differences in the responses to ranibizumab among high myopia patients despite ranibizumab being commonly used to treat this pathology.
This study aims to evaluate the association of VEGFA variants with interindividual differences in the responses to ranibizumab. It also aims to assess the association of genetic polymorphisms previously associated with the response to treatments used in patients with CNV (anti-VEGF drugs and PDT), excluding drug-SNP interactions in non-ophthalmic indications (cancer, hypertension). Furthermore, as a secondary objective, it is intended to characterize these genetic polymorphisms in our population and evaluate their possible association as genetic markers of CNV.

Materials and Methods
A retrospective study including patients with high myopia and CNV treated with ranibizumab between 2014 and 2019 in San Cecilio University Hospital (Granada, Spain). We studied the association of genetic polymorphisms with differences in response to treatment from baseline (BL) up to 1 and 6 months of follow-up.
The Research Ethics Committee of Granada approved the study (0085-N14; 26 May 2014). We obtained written informed consent from all participants and we followed the principles of the Declaration of Helsinki.

Inclusion/Exclusion Criteria
The inclusion criteria were: 1. Diagnosed high myopia and CNV, 2. Treatment with ranibizumab by prescription from ophthalmologists at our hospital, 3. Treated at our hospital, 4. At least six months follow-up with available medical records.
The exclusion criteria were: 1. History of intraocular surgery except for cataract surgery, 2. cataract surgery during the follow-up period, 3. Previous anti-VEGF treatment and 4. Presence of other ocular pathology that may influence the BCVA.
We considered high myopia: AL ≥ 26.00 mm or a spherical equivalent refractive error of ≥−6.0 diopters in phakic eyes and documentation of CNV by optical coherence tomography (OCT) or OCT angiography. If patients underwent bilateral anti-VEGF treatment, both eyes were chosen for analysis. Furthermore, we recruited a control group with high or pathological myopia but with neither CNV nor other associated pathology and not treated with ranibizumab to control the possible association of genetic polymorphisms with the disease and not with the drug response. The inclusion criteria in this group were: 1. Diagnosed high myopia, 2. Age > 18. Exclusion criteria: 1. Another ophthalmic pathology except for cataract surgery, including diagnosis of CNV.

Patients Management
All patients received a 0.5 mg intravitreal dose of ranibizumab. All injections were performed under sterile conditions, and prophylactic topical antibiotics were applied from a few days before to a week after injection. Anti-VEGF therapy was started by a single injection. At baseline and after the first injection, ophthalmologic examinations, including measurement of BCVA, indirect ophthalmoscopy, slit-lamp biomicroscopy with a contact lens by a retina specialist, and OCT were performed at scheduled visits with 1-month intervals. The reinjection criteria included any of the following findings: presence or recurrence of fluid indicative of active CNV on OCT images, new subretinal hemorrhage, and worsening of subjective symptoms, such as metamorphopsia, central scotoma, paracentral scotoma, or VA loss. To recruit both the treatment and control groups, a nurse daily reviewed all the medical records of the patients cited in the ophthalmology unit because of the follow-up of their pathology, and we selected all patients meeting the inclusion/exclusion criteria. At this moment, the nurse asked the patient to sign the informed consent and took four saliva samples with sterile cotton swabs.

Data Management
It was collected the following data from the clinical records of patients: medical record number, age, gender, date of diagnosis, date of initiation of treatment, affected eye, LA, refractive error, location of CNV, baseline VA and VA at 1 and 6 months.
As the primary endpoint of the study, BCVA was measured with the Snellen optotype, in meters, and it was converted to a logarithm of the minimum angle of resolution (logMAR). The increase in logMAR was measured from BL and after 1 and 6 months. The evolution of the patients was ranked, according to the change in logMAR, in IMPROVE-MENT (YES/NO) or WORSENING (YES/NO). IMPROVEMENT was a decrease, of at least 0.1 in the logMAR scale, from BL during the follow-up time (1 or 6 months). On the contrary, WORSENING was an increase, of at least 0.1 in the logMAR scale, during this time.

Procedures for the Inclusion of Genetic Variants in the Study
To include genetic variants that could affect the response that patients have to ranibizumab, we considered including genetic variants that had been previously associated with the response to treatments used in patients with high myopia and CNV (bevacizumab, ranibizumab, aflibercept, and PDT), if this association had been found in patients with any ophthalmic pathology.
To do this, we searched in PharmGKB the following terms: ranibizumab, bevacizumab, aflibercept, and PDT. We then considered for inclusion all the drug-SNP interactions reported as clinical annotations with ophthalmic-related phenotypes, thus excluding those drug-SNP interactions related to phenotypes such as "breast neoplasm" or "hypertension", among others. Since ranibizumab is an anti-VEGF drug, we also considered including any genetic variant in the VEGFA gene previously associated with response to any other drug. To do this, we also searched in PharmGKB the term "VEGFA" (gene), and we considered for inclusion all resulting drug-gene interactions reported as clinical annotations. Table 1 shows all the genetic variants found after this search, thus considered for inclusion in our study. Among these variants, we studied possible linkage disequilibrium. Among those variants that presented some linkage, we chose the one with higher frequency

DNA Extraction and Genotyping
For genotyping, DNA was isolated from saliva using standard procedures. DNA extraction was carried out according to the method by Freeman et al. [34], a non-organic (proteinase K and salting-out) protocol with modifications described by Gomez-Martín A. et al. [35]. The SNPs were genotyped using KASP assay technology (LGC Genomics, Hoddesdon, Hertfordshire, UK) and analyzed with KlusterCaller Software from LGC Genomics.

Statistical Analysis
First, we performed a descriptive analysis of clinical parameters recorded from patients (Table 2). Then we calculated the distribution (number of patients and percentage) of genotypes and MAF of each genetic variant for the total number of patients, control, and study groups (Table 3). Here we also studied the Hardy-Weinberg (H-W) equilibrium of each genetic polymorphism for each group and compared the distribution of genotypes between the control and study groups. This comparison allowed us to assess the association of each SNP with CNV.
As main results, we studied the association between each genetic polymorphism and BCVA (logMAR) improvement or worsening at 1-and 6-months follow-up. We performed an allele comparison analysis and a genotype analysis using each genetic model (recessive, dominant, co-dominant, over-dominant, and log-additive).
For the comparison of the distribution of genotypes between control and study group, and main results, we used the Chi-square. We calculated the odds ratio (OR) and p-values. p-values < 0.05 were considered statistically significant. We also calculated the Akaike information criterion (AIC) and Bayesian information criterion (BIC) for each genetic model and SNP-response association study. We chose for the main results genetic models showing lower AIC and BIC.
The descriptive analysis of clinical parameters, MAFs, genotypic distribution, and its comparison among groups was performed using R commander. The H-W equilibrium analysis, linkage disequilibrium in VEGFA, association study of VEGFA hapotypes with response and main results were conducted using the SNPstats online tool [36].
The sample size calculation was based on previous research articles evaluating the influence of considered genetic polymorphisms on ranibizumab response on patients with other pathologies [27][28][29][30][31][32][33]. Furthermore, we recruited the total number of patients with high myopia and treated them with ranibizumab in our hospital for five years. Both sample size and statistical influence of SNPs on genetic diseases were studied in several works which investigated the impact on pathological phenotypes [37][38][39][40].

Results
In total, between 2014 and 2019 in our hospital, n = 100 patients and n = 113 eyes were diagnosed with CNV secondary to high myopia and treated with ranibizumab, thus eligible for this study. DNA concentration from one patient (n = 1 eye) was not enough to do the needed genotypes. Finally, n = 99 patients and n = 112 eyes were recruited and included in the study, with a mean age of 57.5 ± 13.9 years and 75% women. Two out of three eyes presented juxta-foveal development of CNV, n = 30 (26.8%) sub-foveal and n = 8 (7.1%) extra-foveal location. Almost all eyes (92%) were not previously treated with LP or PDT because of their high myopia, while n = 8 (7.1%) and n = 1 (0.9%) had been treated with LP and PDT respectively. Mean spherical equivalent refractive error was 12.1 ± 5.4, mean BCVA (logMAR) = 0.62 ± 0.48 at baseline; 0.42 ± 0.41 at 1 month and 0.36 ± 0.36 at 6 months follow-up ( Table 2).
In the control group, we recruited n = 116 patients and n = 219 eyes, with mean age 57.5 ± 15.1 years and 68% women. Furthermore, n = 7 eyes could not be assessed because of the following causes: DNA concentration from 4 patients (n = 4 eyes) was not enough to do the needed genotypes. Saliva samples from 2 patients (n = 2 eyes) were lost, and one patient (n = 1 eye) was diagnosed with mCNV one month after being recruited, thus crossed from the control to the treatment group.

Genotypic Distribution, H-W Equilibrium, and Association of Genetic Variants with CNV
All the genetic variants included in the analysis showed MAFs higher than 0.1. Studied SNPs in ARMS2, CFH, and F13A1 showed significant differences neither in the H-W equilibrium analysis nor the comparison of the genotypic distribution between control and study groups (Table 3).
Considering the total patients genotyped (control + study; N = 215), only variants in VEGFA showed a deviation from the H-W equilibrium. If we look at the genotypic distribution of this gene in the control or study groups (Table 3), we see that deviations from the H-W equilibrium happen in the study group (rs699947, rs3025000 and rs2010963) and no significant differences were found in the control group. Furthermore, one of these genetic polymorphisms (rs3025000) showed a p-value = 0.063 in the comparison of genotypic distribution between study and control group, and most of the studied SNPs in this gene (rs25648, rs699947, rs3025000, rs1570360 and rs2010963) were found to be almost full linked among them (p-value < 1 × 10 −5 ; D' > 0.95; r > 0.25) ( Figure S1).
Regarding CXCL8 (rs4073), this SNP showed a deviation from H-W equilibrium (p = 0.01) in the control group with n = 32 (27.6%) individuals carrying A/A genotype (wildtype), n = 43 (37.1%) A/T and n = 41 (35.3%) T/T. We also found a trend for association (p = 0.075) in the comparison between genotypic distribution between control and study group.
Finally, regarding NRP1 (rs2070296), this SNP also showed statistically significant differences in the distribution of genotypes among control and study groups (p = 0.011), but without showing deviations from H-W equilibrium.
As we can see, F13A1 (rs5985), ARMS2 (rs10490924), CFH (rs1061170) and VEGFA (rs3025040) are the only SNPs not showing significant (p > 0.05) or close to significance (p < 0.1) differences about their genotype distribution among control and study groups, or deviations from the H-W.

Allele Association Study with Response
In the allele association study with response, the CFH (rs1061170) was the only SNP showing significant results at 1-month follow-up (Table 4A). The C allele was found to protect against BCVA worsening (OR = 0.13; 95%CI = 0-0.91; p = 0.025) compared to T allele. At 6-months follow-up (Table 4B), we found that ARMS2 (rs10490924) G allele is associated with increased BCVA improvement (OR = 1.95; 95%CI = 0.98-3.83; p = 0.037) and showed a trend for the association with worsening (p = 0.073) compared to the T allele.
Regarding the CFH (rs1061170) in the same follow-up period (6-months), the C allele showed close to significant results for its association with increased BCVA improvement (OR = 1.82; 95%CI = 0.9-3.8; p = 0.075) compared to T allele. Table 4 shows results about the allele association study with response to ranibizumab (BCVA improvement or worsening) in high myopia patients of included genetic polymorphisms. Table 4A shows results at 1-month follow-up and Table 4B at 6-months follow-up.

Genotype Association Study with Response
In the genotype association study with the response at 1-month follow-up (Table 5A), the CFH (rs1061170) was again the only, among studied SNPs, associated with response. The TT genotype showed higher rates of worsening (recessive model: TT vs. CT-CC; OR = 8.83; 95%CI = 1.05-74.3; p = 0.013).
At 6-months follow-up (Table 5B), the genotype's association study with response showed that ARMS2 (rs10490924) GG genotype is associated with increased BCVA improvement (OR = 2.44; 95%CI = 1.05-5.63; p = 0.035) and decreased BCVA worsening (OR = 0.26; 95%CI = 0.08-0.90; p = 0.024) compared to GT or TT genotypes. It also showed that CFH (rs1061170) CC genotype is associated with increased BCVA improvement (p = 0.005) compared to TC or TT genotypes, but we did not find patients carrying CC genotype and non-BCVA improvement. Table 5A,B indicate at 1-and 6-months follow-up respectively, the results of the genotype association study with BCVA improvement/worsening of CFH (rs1061170) and ARMS2 (rs10490924). Supplementary Tables S1 and S2 show these results for the other included SNPs.
Regarding the association study of VEGFA variants with response, we found no significant results (Supplementary Table S3).
Based on AIC and BIC, the dominant and recessive models are respectively the best explaining the association of ARMS2 (rs10490924) and CFH (rs1061170) SNPs with the response to ranibizumab in patients with mCNV.   Summarizing, among our patients and studied SNPs, the CFH (rs1061170) and ARMS2 (rs10490924) were the only associated with the response at 1-month or 6-months follow-up in the allele or genotype association study. The CFH (rs1061170) C allele is associated with protection against BCVA worsening at 1-month and with BCVA improvement at 6-months; TT genotype is associated with worsening at 1-month, and CC genotype with improvement at 6-months. Regarding ARMS2 (rs10490924), the G allele, and GG genotype, are associated with BCVA improvement at 6-months.
Among those other studied genetic polymorphisms, we found no genotypes (see Supplementary Tables S1 and S2) or alleles (Table 4A,B) associated with ranibizumab response in patients with high myopia regardless of the follow-up time.

Discussion
Ranibizumab and bevacizumab are both anti-VEGF drugs used in the treatment of mCNV among other pathologies. The influence of genetic polymorphisms on anti-VEGF drugs response has been widely studied in patients with AMD, especially for bevacizumab [41,42]. In this study, we assessed in patients with mCNV, if ten genetic variants in the VEGFA gene and those genetic polymorphisms previously related to interindividual differences in the response to drugs used in the treatment of CNV (anti-VEGFA drugs and PDT), are associated with ranibizumab efficacy.
We still had limitations in this study. It is a retrospective study; we recruited a control group, but we did not collect data about BCVA in this group since this was only to control the possible association of genetic polymorphisms with CNV, not with ranibizumab response. Also, the study cohort is low and it should be increased in further studies. We recruited only n = 100 patients in 5 years period of recruitment but these were the total patients treated with ranibizumab with high myopia in our hospital. Because of this, CFH (rs1061170) association with ranibizumab response at 6 months could not be confirmed since we did not find at least one patient carrying CC (recessive homozygous) genotype and non-BCVA improvement (Table 5B). In the same regard, this impeded to perform a multivariant analysis considering combined genetic and clinical parameters affecting ranibizumab response. Following in this regard, we included both eyes of patients with bilateral treatment considering that they might progress in a different way depending on the expression/silencing of PGx variants. This is an observational study, and we did not conduct a clinical trial assessing the usefulness of tailoring anti-VEGF treatment depending on PGx. But this would not be ethical since not even these genetic polymorphisms had been related to interindividual differences on ranibizumab response.
Our results support the need for further studies improving these limitations, thus performing a clinical trial including combined PGx, clinical parameters (e.g., previous treatments as LP, PDT), in a multivariant analysis, and in a larger cohort including patients from different populations.
We considered for inclusion in this study only genetic polymorphisms previously associated with response to drugs used in AMD or CNV patients and we did not include in the analysis other SNPs found in those genes if they were not associated in this regard. On the other hand, we found only one SNP, in CXCL8, the rs1126647, previously related to drug response, but this had been associated with sunitinib response [43], not with any treatment used in AMD or CNV. No other SNP found in assessed genes were related to any other drug response when this study was performed.
Also, we calculated BCVA improvement/worsening based on BCVA (logMAR) change from BL to 1 or 6-months follow-up. Results about the association of SNPs with mean logMAR change (quantitative crude parameter) were not considered and this should be addressed in future studies. As we can see in Table 2, mean logMAR at baseline and 6 months follow-up were 0.63 ± 0.48 and 0.36 ± 0.39 respectively, thus interindividual variability about this parameter may lead to confusing results about the association between genetic polymorphisms and response.

ARMS2 (rs10490924) and CFH (rs1061170)
The ARMS2 gene encodes a protein with unclear biological function [44] expressed in the choroid, monocytes, and microglia cells, mediating its activity against apoptotic cells by complement activation [45], so genetic polymorphisms in this gene may lead to decreased apoptotic cell clearance, influencing CNV development [46] and anti-VEGF drugs response as ranibizumab.
This SNP had been previously related to anti-VEGF drugs response with inconclusive results. Ku Kang H. et al. [47] found that patients with the GT and TT genotype had greater improvements in visual acuity at 6 months of treatment with bevacizumab in the east Asian population, as compared to those with the GG genotype, but reporting n = 8 (10.7%) of patients carrying GG genotype, which does not accord with MAFs for this SNP reported in large datasets [48,49]. Also, many studies found no association with response [27,50,51] but considering different follow-up periods, endpoints, and none of them in high myopia patients. On the other hand, a study by Tian J. et al. [31] reported an association between TT genotype with decreased response to bevacizumab in people with Macular Degeneration.
Among our patients, ARMS2 (rs10490924) showed to be associated with ranibizumab response at 6-months follow-up. The GG (dominant homozygous) genotype was related to increased response (greater BCVA improvement and decreased worsening) compared to GT or TT genotypes, G allele was associated with greater BCVA improvement, and showed a p-value = 0.073 in the association study with decreased worsening compared to T allele (minor). We did not find differences about genotypic distribution among treated and control group or deviation from H-W equilibrium. On the other hand, we did not find significant differences at 1-month follow-up.
The CFH gene encodes the complement factor H expression, which inhibits components in the complement cascade related to inflammation and drusen appearance [52]. Changes in the CFH gene may result in a malfunctioning CFH, rendering it unable to inhibit the complement cascade [53], especially its affinity to C-reactive protein (CRP), enhancing CRP levels in the choroid. This may lead to increased levels of VEGF resulting in neovascularization [54], thus genetic polymorphisms in CFH may influence ranibizumab response, especially in CNV patients.
Previous studies found inconclusive results about the association of CFH (rs1061170) with response to anti-VEGF drugs in AMD patients [27,47,51], but no studies were done with high myopia patients.
Among our patients at 1-month follow-up, the CFH (rs1061170) C allele was found to protect against BCVA worsening (OR = 0.13; 95%CI = 0-0.91; p = 0.025) and TT genotype showed higher rates of worsening (recessive model: TT vs. CT-CC; OR = 8.83; 95%CI = 1.05-74.3; p = 0.013). At 6-months the C allele showed a trend for the association with increased BCVA improvement (p = 0.075) compared to T allele and CC genotype associated with increased BCVA improvement (p = 0.005) compared to TC or TT genotypes, but we did not find patients carrying CC genotype and non-BCVA improvement, so we cannot confirm this association.
As we can see, both ARMS2 (rs10490924) and CFH (rs1061170) showed quite similar results at 1 or 6-months but differences. These discordances about the association with response depending on follow-up time might be explained by the lack of response at one month or by the size of the hemorrhage produced by the neovascularization. Some patients had more extensive bleeding at baseline than others. These patients may have a favorable genetic profile but could have a slower response to treatment with a slower recovery of BCVA.

VEGFA Genetic Polymorphisms
VEGFA gene is a member of the PDGF/VEGF growth factor family which encodes a heparin-binding protein. This growth factor induces proliferation and migration of vascular endothelial cells and it is essential for both physiological and pathological angiogenesis.
Genetic variants in VEGFA included in this study were not associated with ranibizumab response among our patients (see Supplementary Tables S1 and S2). Most of them (rs25648, rs699947, rs3025000, rs1570360 and rs2010963) were found to be almost full linked (p-value < 1 × 10 −5 ; D' > 0.95; r > 0.25) (see Supplementary Figure S1), among these five SNPs, three were not in H-W equilibrium (p < 0.05; rs699947, rs3025000 and rs2010963) and remaining two (rs25648 and rs1570360) were almost in disequilibrium (p < 0.1). H-W equilibrium deviations might be explained by linkage disequilibrium.
The only SNP in VEGFA included in this study in the H-W equilibrium, without differences between control and study groups and not linked with others was the VEGFA (rs3025040), but this was neither associated with response to ranibizumab (see Table 4A,B,  and Supplementary Tables S1 and S2).
Anyway, VEGFA variants should be further studied for their associations with drugs responses, characterizing relevant variants, and considering interactions among genetic polymorphisms.

Other Genetic Polymorphisms Included in This Study
CXCL8 encodes a protein member of the CXC chemokine family which acts as a major mediator of the inflammatory response. The encoded protein is commonly referred to as interleukin-8, secreted by mononuclear macrophages, neutrophils, eosinophils, T lymphocytes, epithelial cells, and fibroblasts. Among our patients, the CXCL8 (rs4073) was not associated with ranibizumab response and it showed a deviation from the H-W equilibrium (p = 0.01) and was almost significant (p = 0.075) differences about genotype distribution between control and study groups.
The deviation from H-W equilibrium and genotypic distribution among groups might mean that this SNP is related, in some way, with the absence of CNV, but further studies are needed and other reasonings as the low number of patients carrying this variant might explain this regard.
On the other hand, NRP1 (rs2070296) showed significant differences (p = 0.011) for the genotypic distribution among groups without deviations from H-W equilibrium which suggests its association with CNV development.
F13A1 (rs5985) showed no significant association with response neither in the allele nor genotype analysis, no differences among groups or deviations from H-W equilibrium.

Conclusions
CNV is a complication that may occur in patients with AMD or high myopia. Patients presenting high myopia and CNV are usually treated with anti-VEGF drugs (ranibizumab, bevacizumab, and aflibercept) or PDT. Among AMD patients, some genetic polymorphisms showed differences in the response to these drugs, especially bevacizumab, resulting in different levels of evidence. In contrast, it is difficult to find studies evaluating the influence of these genetic polymorphisms on the response to ranibizumab regardless of the pathology, and no one reports results about patients with high myopia.
Based on our results, ARMS2 (rs10490924) and CFH (rs1061170) SNPs are associated with response to ranibizumab in high myopia patients with CNV.
The ARMS2 (rs10490924) G allele and GG genotype lead to a better response to this drug at 6-months follow-up. By contrast, the CFH (rs1061170) T allele and TT genotype are associated with higher rates of BCVA worsening at 1-month follow-up among these patients.
Finally, included VEGFA genetic polymorphisms are not associated with ranibizumab response in high myopia patients but these might be related to a higher risk of CNV.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to containing clinical and personal information.