Smoking Is a Risk Factor for Pahcychoroid-related Disorders in Asian Patients: a Case Control Study

Background Cigarette smoking has been reported as a risk factor for the development of neovascular age-related macular degeneration (nAMD). However, the associations between cigarette smoking and subtypes of drusen and nAMD were incomplete, as it lacked consideration of pachydrusen or no signicant drusen. Therefore, this study intended to reveal the associations between cigarette smoking and subtypes of drusen and nAMD. Purpose To evaluate the associations between cigarette smoking and subtypes of drusen and nAMD in an Asian population. included in with treatment-naïve including polypoidal and 3 The patients were stratied into never-, former-, and current-smoker groups, and subtypes, including no signicant drusen, soft drusen, subretinal drusenoid deposits (SDDs), and pachydrusen, were analyzed in each group.

prevalence of PCV and type 3 neovascularization is higher and lower in Asians than in Caucasians, respectively [5,6]. These subtypes have different clinical characteristics, and identifying the risk factors for each subtype could reduce the development of nAMD.
Drusen, which are accumulations of extracellular matrix components between the retinal pigment epithelium (RPE) and Bruch's membrane [7], have been described as retinal precursors of AMD [8,9].
Recently, pachydrusen was reported to be thicker choroid-associated drusen [10], and drusen are classi ed into soft drusen, subretinal drusenoid deposits (SDDs), and pachydrusen. This particular type of drusen has been found to be associated with the development of a speci c subtype of nAMD [11][12][13].
Eyes with SDDs have a higher risk of developing type 3 neovascularization. Spaide constructed a novel AMD classi cation system based on three drusen subtypes for the prediction of a more precise prognosis in AMD patients [14].
Of the many risk factors for nAMD, cigarette smoking is consistently associated with the development of nAMD [15][16][17][18][19][20][21]. A population-based cohort study [22] reported that cigarette smoking increased the risk of large soft drusen. However, the drusen classi cation used in previous studies on the association between nAMD and smoking was incomplete, as it lacked consideration of pachydrusen. In addition, those with PCV, a major subtype of nAMD in Asian people, have a comparatively low incidence of drusen [3]. Even among Caucasians, who have a high prevalence of drusen, the incidence of drusen is lower (16.7%) in those with PCV 23 . Therefore, the effect of cigarette smoking on nAMD, especially in Asians, requires the evaluation of the eye without drusen.
To assess the associations between drusen subtypes and cigarette smoking in nAMD patients, we examined the prevalence of no signi cant drusen, soft drusen, SDDs, and pachydrusen in the fellow eyes of newly diagnosed nAMD patients strati ed by smoking history. Moreover, we investigated the 5-year incidence of nAMD in the fellow eyes and the 1-year outcome of anti-vascular endothelial growth factor (VEGF) therapy in the affected eyes in each smoking history group.

Study Design and Participants
We retrospectively evaluated Japanese patients aged 50 years or older with newly diagnosed nAMD, including typical AMD, PCV, and type 3 neovascularization at Kawasaki Medical School between May 2016 and January 2021. The study was complied with the principles of the Declaration of Helsinki. The study was performed with the approval of the Institutional Review Board of Kawasaki Medical School Ethics Committee (2543-1) and is registered in the UMIN Clinical Trials Registry (UMIN000023676). The inclusion criteria were the presence of nAMD diagnosed by funduscopic, swept-source optical coherence tomographic (OCT) (DRI OCT-1 Atlantis; Topcon Corporation, Tokyo, Japan), and angiographic ndings (HRA-2; Heidelberg Engineering GmbH, Dossenheim, Germany), and a best-corrected visual acuity (BCVA) of 20/400 or better at baseline. Data on cigarette smoking were obtained from hospital records and patient recall. The patients were divided into never-, former-, and current-smoker groups. Former and current smokers were those who smoked at least 1 cigarette per day for more than 1 year in their lifetime. Former smokers were those who did not smoke at baseline and had quit smoking for at least 1 year. The patients were treated with a ibercept (Bayer AG, Leverkusen, Germany) for at least one year. Patients who had received or were receiving other anti-VEGF agents (bevacizumab, pegaptanib, ranibizumab) or had undergone laser photocoagulation, vertepor n photodynamic therapy, or submacular surgery were excluded, as were those with choroidal neovascularization (CNV) as a result of high myopia, angioid streaks, or uveitis. Patients with eye diseases that could potentially in uence the clinical features of the studied eyes, such as glaucoma, diabetic retinopathy, or rhegmatogenous retinal detachment, were also excluded.

Group Classi cation
We classi ed drusen subtypes into four groups depending on the condition of the fellow eye as follows: no signi cant drusen, soft drusen, SDDs, or pachydrusen. The type of drusen was determined using fundus color photographs and swept-source OCT according to the criteria presented in a previous study 10 . The no signi cant drusen group included eyes without drusen or eyes with small drusen (size: <63 µm) or few intermediate drusen (number: <20 lesions, size: <125 µm). The soft drusen group included eyes with numerous intermediate drusen (number: ≥20 lesions, size: ≥63 µm and < 125 µm) or one large drusen (size: ≥125 µm) according to the Age-Related Eye Disease Study (AREDS). Eyes with pachydrusen and soft drusen or SDDs were classi ed into the soft drusen or SDDs group, respectively. Eyes with soft drusen and SDDs were classi ed into the SDDs group. The subtype of neovascular AMD (typical AMD, PCV, or type 3 neovascularization) was comprehensively diagnosed on the basis of the ndings of fundoscopy, angiography, OCT, uorescein angiography (FA), and indocyanine green angiography (ICGA).
The diagnosis of PCV was based on ICGA ndings, including polypoidal structures at the borders of the branching choroidal vascular networks. The diagnosis of type 3 neovascularization was based on the characteristic ndings of retinal pigment epithelial detachment with overlying cystic retinal edema on OCT images, intraretinal hemorrhage, and intraretinal vascular anastomoses. Typical nAMD was characterized by the presence of exudative changes due to CNV on FA and ICGA.

Treatment and Assessments
Treatment-naïve nAMD patients received intravitreal a ibercept at Kawasaki Medical School between May 2016 and January 2021. All patients received 3 monthly injections of the anti-VEGF agent initially on a pro re nata basis. If recurrence, including new macular hemorrhage, intraretinal uid, and subretinal uid, was observed, anti-VEGF therapy based on a pro re nata or treat-and-extend regimen was resumed.
All the included patients were followed for 12 months or longer after the initial intravitreal administration of a ibercept and provided written informed consent for treatment with an anti-VEGF agent and participation in the study. The patients underwent a comprehensive ophthalmological examination, including BCVA measurement, slit-lamp biomicroscopy, indirect fundoscopy, fundus color photography, and swept-source OCT. The treatment outcome measures were the change in BCVA and retinal thickness at baseline and 1 year after the start of anti-VEGF therapy, respectively, as well as the number of injections received, the rate of dry macula after the loading dose, and the retreatment-free period after the loading dose. The BCVA was recorded as decimal values, followed by conversion to the logarithm of the minimal angle of resolution (logMAR) units for statistical analysis. Central retinal thickness (CRT) was de ned as the mean retinal thickness measured at the fovea as previously described [24].

Statistical Analysis
The results are presented as means and standard deviations. One-way ANOVA followed by Sidak's test was performed to compare age, BCVA, retinal thickness at the fovea, and the number of injections received among the nAMD-strati ed smoking history groups. The chi-square test followed by residual analysis concerning cross-tabulation was used to compare the proportions of sex, drusen subtype, nAMD subtype, and rate of dry macula after the loading dose among the nAMD-strati ed smoking history groups. Kaplan-Meier analysis was performed to estimate the incidence of nAMD in the fellow eye and the retreatment-free period after the loading dose. Statistical analyses were performed using Ekuseru-Toukei 2012 (Social Survey Research Information Co., Ltd, Tokyo, Japan). P-values < 0.05 were considered statistically signi cant in all analyses. * Chi-square test.

Results
In total, 189 eyes in 189 patients with nAMD were included ( Table 1). The mean patient age was 74.2 ± 8.5 (range, 50-94) years, and there were 58 females and 131 males. All the patients were Japanese. The characteristics of the patients with nAMD in the smoking history groups are shown in Table 1. The analysis included 66, 97, and 26 patients in the never-, former-, and current-smoker groups, respectively.
The three study groups were comparable regarding age (P = 0.15); however, there was a signi cantly higher proportion of males in the former-and current-smoker groups than in the never-smoker group (P < 0.001). The proportions of drusen subtypes in the fellow eyes were signi cantly different (P < 0.001). The proportions of no signi cant drusen and pachydrusen in the fellow eyes were signi cantly higher in the former-and current-smoker groups (P = 0.016 and 0.001), respectively. There were signi cantly higher and lower proportions of SDDs in the fellow eyes in the never-and former-smoker groups (P < 0.001 and < 0.001), respectively. Similarly, the proportions of nAMD subtypes in the affected eyes were signi cantly different (P = 0.022). The proportions of typical AMD and PCV in the affected eyes were signi cantly lower and higher in the current-smoker group (P = 0.005 and 0.041), respectively. There was a signi cantly higher proportion of type 3 neovascularization in the affected eyes in the never-smoker group (P = 0.037).
After a follow-up period of 5 years, 16 of 65 (24.6%) fellow eyes developed nAMD. The numbers of eyes that developed nAMD in the fellow eye were 31.3%, 38.5%, and 5.0% in the never-, former-, and currentsmoker groups, respectively. The Kaplan-Meier curve showed that the incidence of nAMD development in the fellow eye was not signi cantly different among smoking history groups (P = 0.051, Fig. 1). However, nAMD development in the fellow eye was signi cantly associated with the drusen subtype in the fellow eye. The numbers of eyes that developed nAMD in the fellow eye were 21.9%, 16.7%, 83.3%, and 0% in the no signi cant drusen, soft drusen, SDDs, and pachydrusen groups, respectively (P = 0.0028, Fig. 2). The incidence of nAMD development in the fellow eye was signi cantly higher in the SDDs group. Therefore, we evaluated the relationship between smoking history and the development of nAMD in the fellow eye according to each drusen subtype. The SDDs and pachydrusen groups were excluded due to their small sample sizes (n = 6 and 3, respectively); the numbers of patients who developed nAMD in the fellow eye were 7 of 32 in the no signi cant drusen group and 4 of 24 in the soft drusen group. The incidence of nAMD development was not signi cantly different among smoking history groups in either group (no signi cant drusen group: P = 0.97, soft drusen group: P = 0.82, Fig. 3).
To evaluate the effect of smoking on anti-VEGF therapy, we analyzed the data of nAMD patients treated with a ibercept who completed 12 months of follow-up. A total of 147 eyes were eligible for analysis. The outcome of anti-VEGF therapy in patients with nAMD is shown in Table 2 (55, 72, and 20 eyes in the never-, former-, and current-smoker groups, respectively). The association between age and smoking history was not statistically signi cant (P = 0.11). The mean BCVA and CRT in the overall sample improved from 0.29 and 320.0 µm to 0.13 and 232.2 µm over 12 months, respectively; there were no signi cant differences among smoking history groups. Additionally, we evaluated the injection number, the presence of residual intraretinal or subretinal uid after the 3-month loading dose, and the retreatment-free period after the loading dose to investigate the treatment frequency. The mean number of injections, rate of dry macula after the loading dose, and retreatment-free period after the loading dose in the overall sample were 6.5, 81.0%, and 3.3 months, respectively; there were no signi cant differences among smoking history groups. Similarly, the Kaplan-Meier curves for the retreatment-free periods after the loading dose were not signi cantly different among smoking history groups (P = 0.30, Fig. 4).

Discussion
In our study, ever smokers (former-and current-smokers) had signi cantly higher proportions of fellow eyes with no signi cant drusen or pachydrusen and affected eyes with PCV than never-smokers, suggesting that cigarette smoking was associated with pachychoroid-related abnormalities. "Pachychoroid" is a term to de ne macular diseases characterized by a thick choroid, choroidal vascular hyperpermeability, and dilatation of the choroidal vessels in Haller's layer with attenuation of inner choroidal vessels, including choriocapillaris and Sattler's layer [25]. Although it is ambiguous whether the attenuation of inner choroidal vessels is a primary pathologic change or a secondary morphological change, inner choroidal attenuation could cause ischemic damage to the RPE [26, 27], followed by CNV, including pachychoroid neovasculopathy (PNV) [28] and PCV [29]. Miyake et al. evaluated differences in genetic backgrounds and clinical features between PNV and non-PNV of Japanese patients with nAMD and demonstrated that PNV patients were signi cantly younger and had lower genetic risk scores than non-PNV patients [30]. In additon, the frequency of genes associated with nAMD (CFH rs800292) in PNV patients was comparable to that in normal Japanese subjects. They suggested that the risk factors for pachychoroid-dependent CNV may differ from those for drusen-dependent nAMD. However, the risk factors for the development of pachychoroid-dependent CNV have not been clari ed. Cigarette smoking is a known risk factor for traditional nAMD. Cigarette smoke contains more than 4000 chemicals [31] that cause oxidative damage and ischemia in almost all organs in the body and induce many chronic diseases, such as cerebral stroke, myocardial infarction, and chronic obstructive pulmonary disease. The mechanism of CNV induced by smoking has not been fully elucidated; however, one pathway by which smoking accelerates the development of CNV is impairment of the RPE and Bruch's membrane through the accumulation of reactive oxygen species induced by components of cigarettes and ischemia due to reduced blood ow in the choriocapillaris. Thus, our results strongly indicate that smoking, which causes ischemic damage to the RPE, plays an essential role in the development of pachychoroid-related abnormalities.
The proportions of fellow eyes with SDDs and affected eyes with type 3 neovascularization were higher in the never-smoker group. Type 3 neovascularization [32], also known as retinal angiomatous proliferation, is characterized by macular neovascularization originating from the retinal vessels, unlike PCV and typical AMD. Yannuzzi et al. reported that type 3 neovascularization tended to affect a higher proportion of females than males, with a sex ratio of more than 2:1 [33]. Even though Asian nAMD patients are predominantly male, Asian patients with type 3 neovascularization are predominantly female [3]. Several epidemiologic studies performed in Asia documented a higher prevalence of late AMD in males than in females, which is attributed to the substantially higher prevalence of smoking in males than in females in Asian countries [34][35][36]. These results suggest that the effect of cigarette smoking on the development of type 3 neovascularization is less than expected. More than 90% of eyes with type 3 neovascularization have SDDs [37], and SDDs are important in the pathophysiology of type 3 neovascularization. The mechanism of SDD formation is considered to be associated with choroidal circulatory disturbances [38][39][40], lipid transport abnormalities [41][42][43], or retinoid cycle disorers [44,45]. deposits and large choroidal vessels or the stroma [47]. In the SDD group in our study, the proportion of never-smokers was 72.2%, which is comparable to the rate in the healthy elderly population (70.0% [112/160], data not shown), and our results supported the lack of association between SDDs and choroidal vessels. Although further research on the association between cigarette smoking and the development of SDDs or type 3 neovascularization is required, the effects of cigarette smoking on each drusen subtype likely differ.
The 5-year incidence of nAMD in the fellow eyes was 24.6% in this study. A recent study in Asia that investigated the 5-year progression rate of nAMD in the fellow eye reported a rate of 20.9% [48]. They classi ed nAMD patients using the new drusen classi cation and showed that eyes with soft drusen and/or SDDs had a signi cantly higher risk of nAMD in the fellow eye at 5 years than eyes with pachydrusen or no signi cant drusen. Another study also demonstrated that the pachydrusen group had a signi cantly lower frequency of nAMD development in the fellow eye than the soft drusen and SDD groups [49]. In this study, the rate of development of nAMD in the fellow eye did not differ among smoking history groups; however, there was a signi cant difference among drusen subtypes, similar to previous studies. These results suggest that the new drusen classi cation could play an important role in predicting the prognosis of patients.
The limitations of this study include its retrospective study, single-center design and relatively small number of nAMD patients. All participants in this study were Japanese. Since the frequency of nAMD subtypes varies among racial groups, the results may not be generalizable to other racial and ethnic groups. We showed that cigarette smoking did not affect the 1-year outcomes of anti-VEGF therapy, but we did not con rm long-term outcomes. The Comparison of Age-related Macular Degeneration Treatments Trials (CATT) showed that cigarette smoking did not in uence on visual prognosis at 1 year after initial treatment in 1,105 nAMD patients who were treated with ranibizumab or bevacizumab [50]. However, the ve-year outcomes in the CATT study reported that cigarette smoking was associated with an increased risk of poor visual prognosis (20/200 or worse) [51]. In a prospective study examining 987 eyes, nonsmokers had higher visual acuity improvements than former and current smokers [52]. In contrast, McKibbin et al. demonstrated that the visual outcome associated with anti-VEGF therapy for nAMD was not signi cantly different among never smokers, former smokers, and current smokers [53]. This inconsistency may be due to differences in follow-up periods, agents, treatment methods, race, nAMD subtype ratios, and smoking history classi cation. It should be noted that the patients in this study were treated with a ibercept monotherapy, and the study did not include patients treated with other anti-VEGF agents or photodynamic therapy.
In conclusion, ever smokers had a signi cantly higher proportion of pachychoroid-related abnormalities (no signi cant drusen or pachydrusen in the fellow eyes and PCV in the affected eyes) that nonsmokers, suggesting that cigarette smoking could promote the development of pahychoroid diseases due to ischemic damage in the RPE. Never smokers had a signi cantly higher frequency of SDDs in fellow eyes and type 3 neovascularization in affected eyes. Therefore, circulatory insu ciency may not be strongly associated with the development of SDDs or type 3 neovascularization. Cigarette smoking is a wellknown modi able risk factor for nAMD development, and the new drusen classi cation allowed us to clear differences in the contribution of cigarette smoking for nAMD development. Our present ndings should heighten awareness of the risks of cigarette smoking in both the general population and health professionals. Availability of data and materials Data will be made available upon reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This study received no funding support.  Tables   Due to technical limitations, table 1 and 2 tif are only available as a download in the Supplemental Files section. Figure 1 Five-year incidence of nAMD in the fellow eyes strati ed by smoking history. The numbers of patients who developed nAMD in the fellow eye were 31.3%, 38.5%, and 5.0% in the never-, former-, and currentsmoker groups, respectively. The Kaplan-Meier curve demonstrated that the incidence of nAMD development in the fellow eye was not signi cantly different among the three groups (P = 0.051).

Figure 2
Five-year incidence of nAMD in the fellow eyes classi ed drusen subtype. The numbers of patients who developed nAMD in the fellow eye were 21.9%, 16.7%, 83.3%, and 0% in the no signi cant drusen, soft drusen, SDDs, and pachydrusen groups, respectively. The incidence nAMD development in the fellow eye was signi cantly associated with the drusen subtype in the fellow eye (P = 0.0028) Figure 3 Five-year incidence of nAMD in the fellow eyes strati ed by smoking history in soft drusen and no signi cant drusen groups. The incidence of nAMD development was not signi cantly different among smoking history groups in either group (no signi cant drusen group: P = 0.97, soft drusen group: P = 0.82).

Figure 4
Retreatment-free periods after the loading dose. The Kaplan-Meier curves for the retreatment-free periods after the loading dose were not signi cantly different among smoking history groups (P = 0.30).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.