Peripapillary choroidal thickness after intravitreal ranibizumab injections in eyes with neovascular age-related macular degeneration

Age-related macular degeneration (AMD) is a leading cause of vision loss in the elderly worldwide [1]. With the development and introduction of anti-vascular endothelial growth factor (VEGF) therapy, the functional outcomes of neovascular AMD have been markedly improved [2‐4]. Ranibizumab (Lucentis; Genentech, Inc., San Francisco, CA; co-developed by Genentech, Inc and Novartis), is an FDA approved VEGF-A inhibitor that is commonly used for the treatment of AMD [5].

Previous studies have suggested that AMD may be a vascular disorder associated with altered blood flow in the choroid [6, 7]. Because abnormalities in choroidal circulation and structure are thought to be associated with the pathogenesis of AMD [8, 9], there have been many studies on choroidal thickness (CT) in AMD [10‐13]. Recently, some studies have suggested that the CT in patients with neovascular AMD may be affected by intravitreal injection of ranibizumab (IVR) [14, 15]. Because VEGF is known to be associated with the maintenance of choriocapillaries and ocular blood flow, there have been concerns regarding potential undesired effects of anti-VEGF therapy on the choroid [16, 17]. However, recent studies found that CT changes following IVR treatment for neovascular AMD were limited to the choroid of the macular region [14, 15]. It is not certain whether CT outside the macula is affected by IVR, and the effect of anti-VEGF therapy on CT of normal choroid without overlying retinal pathology remains unclear [18].

The choroid is a highly vascular tissue of the eye that supplies blood to the outer retina, retinal pigment epithelium (RPE), and possibly the optic nerve head [19]. The choroid is supplied by various ciliary arteries that originate from the ophthalmic arteries [20]. The branches of the posterior ciliary artery (PCA) consist of long PCAs and short PCAs [20]. The posterior choriocapillaries in the peripapillary and submacular region are supplied by the short PCAs. These arteries have strictly segmental blood flow without anastomosis and supply a well-defined sector of the choroid [21]. The medial PCA supplies the nasal choroid, while the lateral PCA supplies the area of the choroid not supplied by the medial PCA [20]. Therefore, blood flow to the nasal peripapillary choroid could differ from the flow to the temporal peripapillary and subfoveal choroid. Recently, several studies on the choroid outside of the macula in retinal disease have focused on the peripapillary choroid and the nasal peripapillary choroid represented the choroid outside of macula [22‐24]. We hypothesized that the change in CT after IVR would differ between the macula and areas outside of the macula. We also investigated the peripapillary and subfoveal CT in neovascular AMD and the change in CT after IVR treatment.

The Institutional Review Board of Korea University Medical center approved this study (IRB No. ED14121), and all research and data collection was conducted in accordance with the tenets of the Declaration of Helsinki. We performed a retrospective case–control study of patients 50 years of age or older who were diagnosed with neovascular AMD between February 2011 and March 2013 at Korea University Medical Center. We included patients who underwent IVR for neovascular AMD. Age-matched controls were selected from the spectral domain optical coherence tomography (SD-OCT) database based on classification of normal eyes. Patients with a primary diagnosis of AMD, a history of laser photocoagulation of the macula, serous detachment of the macula, rhegmatogenous or tractional retinal detachment, high myopia, tapetoretinal degeneration, optic disc abnormalities such as glaucoma, beta zone peripapillary atrophy, or neurologic deficits were excluded from the control group.

Neovascular AMD was diagnosed in patients who were older than 50 and showed exudative changes from choroidal neovascularization (CNV) on fluorescein angiography (FA) and SD-OCT. This study excluded cases of polypoidal choroidal vasculopathy (PCV). PCV was diagnosed based on the presence of CNV on FA with characteristic findings on SD-OCT and if any one of the following criteria were met: 1) branching vascular networks underneath the RPE and focal hyperfluorescent polyps were present on indocyanine green angiography, 2) recurrent hemorrhagic or serous pigmented epithelial detachments were detected, or 3) the existence of an elevated reddish-orange lesion protruding from the choroid on fundus examination [25‐28]. Neovascular lesions were classified into type 1 (sub-RPE), type 2 (subretinal) or type 3 (intraretinal) according to the MPS criteria and to the guidelines provided by Freund and colleagues based on angiographic and OCT features [29‐31]. Classifications were independently determined by two retina specialists (C.Y. and J.A.). Disagreements were resolved upon further review by two observers. Cases with refractive errors ≤ -6.0 diopters or axial length ≥ 26.0 mm and cases with a history of cataract surgery within the past 6 months, refractive surgery or vitreoretinal surgery were also excluded from this study. Patients with a history of glaucoma, optic nerve disorder, beta zone peripapillary atrophy, retinal or choroidal disorder including vascular disease or uveitis and previous treatment with photodynamic therapy or other intravitreal anti-VEGF injection were also excluded from this study. Images were excluded if they were found to be low quality with a signal strength indicator (Q-factor) value < 45 or if the choroidal layer was not visible due to severe subfoveal hemorrhage. Data were collected at baseline, 3 months and 6 months after IVR.

After reviewing the medical records of 79 patients who were diagnosed with neovascular AMD, we excluded 23 cases of PCV, two cases of high myopia, one case of history of vitreoretinal surgery, three cases of low quality image, six cases of glaucoma, five cases of beta zone peripapillary atrophy. In a total of 39 eyes of 39 patients with neovascular AMD and an equal number of controls eyes were included in this study. Mean age at diagnosis in the neovascular AMD group was not different from that of the control group (Table 1). There were no differences in sex or lens status between the two groups. The mean number of IVRs within the 6-month period was 4.4 ± 0.9 (range, 3–6). Mean BCVA (LogMAR) of the 39 patients was 0.73 ± 0.45 at baseline, 0.50 ± 0.38 at 3 months, and 0.51 ± 0.44 at 6 months. Mean BCVA values at 3 months and 6 months were significantly improved when compared to baseline (all, P < 0.001). At baseline, the neovascular lesions were categorized as type 2 (35 eyes, 89.7 %) and type 3 (4 eyes, 10.3 %).

In this study, macular and peripapillary CT in the AMD group did not differ from those in the control group at baseline. The macular CT results in AMD patients and normal controls are consistent with those found in previous studies, which revealed no significant differences between AMD eyes and normal eyes [11‐13]. Because it has been suggested that AMD is a vascular disorder associated with altered blood flow in the choroid [6, 7], we wondered if there were regional differences in CT between the macula and outside of the macula. However, we found no such difference in CT between the macula and outside of the macula in the neovascular AMD and control groups.

Peripapillary CT outside of the macula did not change after IVR at 6 months. These findings differed from the subfoveal choroid results. In this study, both subfoveal CT and temporal peripapillary CT decreased after three monthly IVRs, and the reduction in CT was sustained for 6 months. CT outside of the macula, including superior, nasal and inferior peripapillary CT, was not affected by IVR. Even though there was a trend toward a reduction in PCT in all sectors, these differences were not statistically significant. This may indicate that ranibizumab affects the choroid under and around the macula but not the choroid outside of the macula. We did not investigate the change in CT in AMD eyes without treatment. Therefore, the results of our study could not confirm if choroidal thinning of the subfoveal area is directly induced by the IVR or by the disease course of AMD. However, the results of the current study may suggest that the CT outside of the macula is not involved in the change in CT after IVR in eyes with AMD.

In the current study, subfoveal CT decreased after IVR, a result consistent with previous studies [14, 15]. Yamazaki et al. suggested that IVR may have a pharmacologic effect on the choroid underneath the CNV lesion [15]. Changes in CT associated with IVR may be explained by two possible mechanisms. First, decreased intraocular VEGF after IVR might affect the CT. VEGF is known to be associated with vasculogenesis and blood vessel hyperpermeability [35‐37]. Hypoxia and ischemia of RPE from inadequate choroidal perfusion might produce VEGF, which may lead to neovascular AMD [38]. Both the subfoveal choroid and the peripapillary choroid may be affected by increased VEGF. Although the natural course of CT in neovascular AMD is unknown, vasodilatation may be expected in the choroid of neovascular AMD eyes, as VEGF is associated with vascular permeability [36, 37]. After a decrease in intraocular VEGF levels due to IVR, changes in vascular permeability and CT in both the macular and peripapillary region may follow. In this study, the changes in macular and temporal peripapillary CT were significant while CT outside of macula showed no significant changes. This might suggest that IVR mainly affects the abnormal CNV and the hyperpermeability associated with the CNV under and around the macula. In addition, vasoconstriction of the choroid after IVR may affect CT. Although the effect of VEGF on choroidal structure, thickness and large vessels is not clearly defined, VEGF is known to be essential for the development of the choroid and maintenance of the choriocapillaries [39, 40]. Furthermore, the vasoconstrictive effect of ranibizumab might affect the choroidal vessels and thickness [41, 42]. In this study, a significant change in CT was observed only in the macular and temporal peripapillary areas. Thus, even though we cannot exclude the vasoconstrictive effect of IVR on the macula, IVR does not result in a change of CT outside of the macula.

In this study, decreased proportions of subfoveal and temporal peripapillary CT were associated with treatment response after three monthly doses of ranibizumab. However, in multivariate analysis, a decreased proportion of subfoveal CT was a factor associated with treatment response, which might result from the positive correlation between a decreased proportion of subfoveal CT and temporal peripapillary CT. This may suggest that decreased hyperpermeability or vasoconstriction and regression of CNV after IVR change the choroid under and around the macula. However, we cannot conclude whether decreased CT is a preceding factor associated with treatment response or if decreased CT is a result of treatment response, as the decreased ratio of CT is not a preoperative finding but rather a parameter that is calculated from CT before and after IVR.

However, the observed decrease in CT is insufficient to limit the clinical use of ranibizumab. In the present study, although macular and temporal peripapillary CT were decreased at 3 and 6 months after IVR when compared to the initial CT, the reduction in CT was relatively small, and was not associated with a loss of BCVA. In addition, CT after IVR at 3 and 6 months was not different from that of the normal controls. In addition, the non-significant changes in CT outside of the macula suggest that the effect of IVRs outside of the macula is minimal. IVRs decreased temporal peripapillary RT, but may be associated with a decrease in CRT. The RT of other peripapillary areas and RNFL thickness were not affected by IVRs.

In the control group, the subfoveal CT and peripapillary CT showed a negative relationship with age. Previous studies have demonstrated that the subfoveal CT of normal controls and patients with dry AMD are correlated with age, but the eyes with neovascular AMD demonstrated a weaker correlation between subfoveal CT and age [13, 34]. The peripapillary CT was also related to age and had a negative correlation with the results from our previously reported study [33]. In this study, the normal control group showed a negative correlation between subfoveal and peripapillary CT and age, but the neovascular AMD group showed no correlation between subfoveal and peripapillary CT and age. This absent relationship between subfoveal CT and age might suggest an alteration in the choroid of the macula in neovascular AMD and may provide additional insight into the pathogenesis of neovascular AMD [15].

This study has several limitations. First, because it is a retrospective study with short-term results and contains a small number of cases, our study may have limited generalizability. We could not see a short-term change in CT 1 month after IVR, and it is insufficient to draw conclusions on the long-term effect of IVR on the choroid based on the short follow-up period. Further studies with larger sample sizes and longer study periods are needed. Second, because of the retrospective nature of this study, diurnal variation of CT was not considered. Mean diurnal variation of subfoveal CT in healthy individuals is about 33.7 μm and the diurnal variation of the peripapillary CT remains unclear [43]. The changes in the mean macular and peripapillary CT in this study were relatively small, and the changes in macular CT were less than those expected from diurnal variation. Although most patients visited the outpatient clinic during the day, this study did not adjust for diurnal changes in CT. Third, we could not obtain information about refractive error and axial length in all patients. Two patients in the AMD group and four patients in the control group with pseudophakia had no information available on refractive error, although the axial length in all of these patients was less than 26.0 mm. Because axial length and refractive error are known to be associated with subfoveal CT and peripapillary CT, the lack of this information in the analysis is a shortcoming of this study. Fourth, although CT measurements were performed by two independent examiners, the method was manual as no automated method exists. Fifth, we only matched patients by age, but not by sex. Although there were no differences in these characteristics between the two patient groups, the results may still be affected. Sixth, we only measured CT of the peripapillary or macular region. It could not represent CT beyond the posterior pole. Finally, because the treatment was based on monthly dosing and a PRN schedule, the number of injections received at 6 months differed among patients. However, there was no significant correlation between the difference in subfoveal or peripapillary CT at 6 months and the number of injections.

The mean peripapillary CT of the AMD group was not significantly different from that of the age-matched controls. Apart from the CT in the macular area, the mean peripapillary CT and peripapillary CT outside of the macula did not change 6 months after IVR. This result may suggest that the change in CT after IVR may be limited to the macula in eyes with neovascular AMD.