Introduced by Pang and Freund [
1] in 2015, pachychoroid neovasculopathy (PNV) is a clinical entity characterized by type 1 choroidal neovascularization, focal choroidal thickening, and dilated choroidal vessels. The pachychoroid condition has been increasingly associated with diverse macular disorders, including those without exudative changes, such as pachychoroid pigment epitheliopathy (PPE), and those with exudative changes, such as central serous chorioretinopathy (CSC), PNV, and polypoidal choroidal vasculopathy (PCV) [
2]. PNV can be distinguished from neovascular age-related macular degeneration (AMD) based on the absence of drusen, the presence of pachydrusen, choroidal thickening, and younger onset age, indicating a distinct underlying pathogenesis [
3].
The spectrum of pachychoroid disorders has recently been proposed as a pathophysiological continuum [
4]. In this continuum, non-neovascular early stages can progress to late neovascular stages [
4]. Consequently, PNV falls within the pachychoroid spectrum, which also includes CSC, PPE, and PCV [
1,
5]. Although these diseases share characteristic choroidal alterations, known as pachychoroid, each remains a distinct entity.
Despite most CSC cases resolving spontaneously within 3 months, approximately 10% of patients could experience multiple recurrences or progress to a chronic course [
6]. Chronic CSC is characterized by multifocal, irregularly distributed changes in the retinal pigment epithelium (RPE), photoreceptor damage, and persistent subretinal fluid (SRF). Moreover, PNV can develop as a late neovascular complication of chronic CSC in some cases [
7]. After an average disease duration of 17 years for CSC, the prevalence of PNV has been reported to be as high as 24% [
8].
Previous literature suggests that CSC complicated by PNV is associated with poor visual acuity and low rates of complete fluid resolution [
9]. However, the current literature lacks data comparing treatment patterns and long-term visual outcomes between patients with PNV at initial evaluation and those with new-onset PNV during follow- up. This study aimed to analyze the clinical characteristics and outcomes of chronic CSC, whether presenting with initial PNV, new-onset PNV, or without PNV.
Materials and Methods
Ethics statement
This single-center, retrospective study was conducted in accordance with the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board of Yonsei University College of Medicine (No. 4-2023-1633). The requirement for written informed consent was waived due to the retrospective study and the deidentification of patient data.
Participants
Medical records of patients aged 20 to 55 years with diagnoses of chronic CSC from January 1, 2009, to December 31, 2022, were retrieved from the electronic medical database. CSC was defined as the presence of SRF on clinical examination and detected through spectral domain optical coherence tomography (OCT) using Spectralis OCT (Heidelberg Retina Angiograph 2, Heidelberg Engineering), without other possible causes of SRF. Consistent with previous literature, chronic CSC was defined as having a documented disease duration of at least 6 months, with clinical features such as SRF and RPE changes in the macular area identified through fluorescein angiography (FAG) or OCT imaging [
10]. Patients were enrolled at the time of their first diagnosis of chronic CSC in this tertiary hospital and were censored at the time of death, the last follow-up date, or June 30, 2023, whichever came first. The exclusion criteria were as follows: (1) previous treatment with intravitreal anti-vascular endothelial growth factor (anti-VEGF) injections, photodynamic therapy (PDT), or laser treatments; (2) comorbid vision-threatening disorders, including vitreous hemorrhage, corneal opacity, anterior or posterior uveitis, glaucoma, diabetic retinopathy, and AMD; (3) <6 months of follow-up or follow-up loss before the first resolution of SRF; and (4) absence of FAG, indocyanine green angiography (ICGA), and/or OCT angiography (OCTA) examinations within 1 month from the index date.
Patient demographics, duration of each SRF episode, treatment details, recurrence, and complication data were recorded. The best-corrected visual acuity (BCVA) was measured using Snellen distant visual acuity chart and later converted to the logarithm of the minimum angle of resolution scale. Subfoveal choroidal thickness was measured by drawing a perpendicular line on the outer margin of the RPE using a screen caliper, extending to the inner boundary of the sclero-choroid junction. Group 1 included eyes exhibiting SRF with pachychoroid features and type 1 choroidal neovascularization (CNV), confirmed through initial FAG/ICGA and/or OCTA. Herein, we have used the term “initial PNV with SRF” for such conditions, as it is challenging to differentiate PNV resulting from CSC with isolated PNV without preceding CSC. Among eyes without initial CNV, those that developed type 1 CNV during follow-up were categorized as group 2 (i.e., new-onset PNV), while eyes that did not show type 1 CNV until the end of follow-up were categorized as group 3.
Descriptive statistics were used to characterize baseline characteristics and comorbidities. Continuous variables are expressed as mean ± standard deviation, and categorical variables are reported as frequencies (percentages). Group differences were assessed using the Student t-test or the Mann-Whitney U-test, as appropriate. All tests were two-tailed, with p < 0.05 considered statistically significant. Statistical analyses were conducted using IBM SPSS ver. 23.0 (IBM Corp) and MedCalc ver. 14.8.1 (MedCalc Software).
Results
A total of 144 eyes of 141 patients, aged 44.29 ± 7.00 years, were enrolled in this study. Over an average follow-up period of 82.62 ± 207.71 months, new-onset PNV occurred in 14 eyes (11.8%) of patients with chronic CSC without initial PNV. The baseline characteristics of the study population for each group are provided in
Table 1. Initial SRF height was greater in group 2 than in group 1, but this difference was not statistically significant (135.6 ± 105.1 μm vs. 187.6 ± 140.1 μm,
p = 0.06). No significant differences were identified among age, sex, initial BCVA, systemic comorbidities, and OCT findings of the three groups at baseline (
Table 1).
In group 1 (n = 25), the initial treatment included intravitreal injection in 20 patients (80.0%, average of 9.53 ± 11.53 days after diagnosis), PDT in two patients (8.0%, average of 23.50 ± 14.85 days after diagnosis), and focal laser in one patient (4.0%, the day of diagnosis). Additionally, two patients (8.0%) experienced SRF resolution with conservative care within 2 weeks.
In group 2 (n = 14), the initial treatment included intravitreal injection in seven patients (50.0%, average of 47.0 ± 64.9 days after diagnosis), PDT in one patient (7.1%, 5 days after diagnosis), and focal laser in two patients (14.3%, the day of diagnosis). Additionally, four patients (28.6%) received only conservative care.
In group 3 (n = 105), the initial treatment included intravitreal injections in 51 patients (48.6%, average of 63.88 ± 163.26 days after diagnosis), PDT in 16 patients (15.2%, average of 84.13 ± 83.18 days after diagnosis), and focal laser in 19 patients (18.1%, average of 14.84 ± 19.86 days after diagnosis). Additionally, 19 patients (18.1%) received only conservative care.
The proportion of patients receiving intravitreal injections as the initial treatment was highest in group 1 (group 1 vs. group 2 vs. group 3, 80.0% vs. 50.0% vs. 48.6%; p = 0.017). Among patients who received treatment, treatment initiation tended to occur most quickly in group 1 (9.91 ± 12.11 days vs. 33.40 ± 57.35 days vs. 56.81 ± 132.46 days after diagnosis), although this difference was not statistically significant (p = 0.210).
Table 2 presents details on the time and treatments required to achieve the first complete resolution of SRF. Throughout the follow-up period, 97.22% of patients experienced complete resolution of SRF. The Kaplan-Meier curves depicted in
Fig. 1A illustrate the cumulative incidence of SRF resolution for each group, showing no significant differences in the average time to first resolution among the three groups (11.13 ± 10.70 months vs. 8.14 ± 7.90 months vs. 7.32 ± 9.55 months). The number of anti- VEGF injections required until the first resolution was highest in group 1, although the difference was not statistically significant.
Recurrence of SRF occurred in nine eyes (64.3%) in group 2, significantly more than the seven eyes (29.2%) in group 1 (
p = 0.04) or 28 eyes (26.7%) in group 3 (
p < 0.01) (
Table 3). The Kaplan-Meier curves for the cumulative incidence of SRF recurrence also indicated a higher recurrence rate in group 2 compared with that in group 1 (
p = 0.06) or group 3 (
p = 0.04) (
Fig. 1B). The time gap between the first resolution and SRF recurrence was similar among the three groups (39.9 ± 47.6 months vs. 32.6 ± 34.6 months vs. 29.4 ± 32.6 months). The time taken from the date of recurrence to the date of the second resolution was also similar between the three groups (
Table 3).
The hazard ratios associated with the development of new-onset PNV according to baseline characteristics are shown in
Table 4. Age, systemic comorbidities, initial choroidal thickness, and double-layer signs on the initial OCT showed no significant associations with the development of new-onset PNV.
Table 5 presents the BCVA, choroidal thickness, and SRF height at the initial and most recent follow-ups. Significant improvements in BCVA after treatment were observed in group 1 (0.24 ± 0.27 at the initial follow-up vs. 0.11 ± 0.22 at the most recent follow-up,
p < 0.01) and group 3 (0.26 ± 0.31 at the initial follow-up vs. 0.14 ± 0.22 at the most recent follow-up,
p < 0.01), whereas group 2 did not show significant improvement (0.21 ± 0.16 at the initial follow-up vs. 0.14 ± 0.14 at the most recent follow-up,
p = 0.19). Choroidal thickness was similar pretreatment and posttreatment in all three groups. Although the SRF height significantly decreased at the most recent follow-ups compared with that at baseline in all three groups, the SRF height in group 2 at the most recent follow-up was still significantly larger than that in group 3 (58.3 ± 68.4 vs. 13.2 ± 31.8,
p = 0.02).
Fig. 2 illustrates the clinical course of a patient with initial PNV and accompanying SRF (group 1). Initial OCT revealed SRF and a thickened choroid (
Fig. 2A), and initial FAG/ICG showed PNV (
Fig. 2B, 2C) in the patient’s left eye. Following half-dose PDT, the complete resolution of SRF was evident at 1 month postoperation (
Fig. 2D). However, after 39 months of resolution, the patient experienced visual discomfort following an intra-articular steroid injection, and OCT confirmed the recurrence of SRF (
Fig. 2E). After two intravitreous injections of bevacizumab and one half-dose PDT, OCT showed dry macula but persisting PNV.
Fig. 3 shows the clinical course of a patient with chronic CSC with new-onset PNV (group 2). Initial OCT revealed SRF with shallow irregular pigment epithelial detachment and a thickened subfoveal choroid (
Fig. 3A). Although the initial FAG/ICG showed juxtafoveal leakage in the early phase, no PNV was evident (
Fig. 3B, 3C). Despite half-dose PDT, OCT at postoperative 2 months showed a decreased but persisting SRF (
Fig. 3D). Although minimal SRF persisted, no additional treatment was administered to the patient’s right eye. Subsequent OCT evaluations revealed increased SRF 19 months postoperatively (
Fig. 3E), and FAG revealed PNV with suspicious polyps, suggesting PCV (
Fig. 3F, 3G). Following five intravitreous aflibercept injections, OCT at 11 months postrecurrence revealed the complete absorption of SRF (
Fig. 3H). However, the recurrence of SRF occurred after 5 months.
Fig. 4 depicts the clinical course of a patient with chronic CSC without PNV (group 3). Initial OCT showed SRF and a thickened choroid in the patient’s left eye (
Fig. 4A), whereas FAG/ICG revealed no PNV (
Fig. 4B, 4C). Following a single intravitreous injection of bevacizumab, OCT 6 weeks later revealed persisting SRF (
Fig. 4D). Half-dose PDT was administered, and OCT scans at 1 month post-PDT demonstrated a dry macula (
Fig. 4E). Subsequently, over a follow-up period of 7 years, no SRF recurrence was noted.
Discussion
In this study, although the initial PNV group was administered the highest number of injections or had the longest average time for the initial resolution of SRF, this trend did not reach statistical significance. However, the recurrence rate of SRF was significantly higher in patients with new-onset PNV, compared with that in patients with initial PNV or chronic CSC without PNV. All three groups showed no significant differences in anatomical outcomes at the final visit, although the new-onset PNV group exhibited a larger amount of persisting SRF, but without statistical significance. Although the groups with initial PNV and without PNV achieved statistically significant improvements in corrected visual acuity at the final follow-up, compared with that at the initial evaluation, no significant improvements were observed in the new-onset PNV group.
CSC has recently been categorized within the spectrum of pachychoroid disorders, a group that also encompasses conditions such as PPE, PNV, and PCV [
11]. Recognizing these pachychoroid spectrum diseases is crucial because they can mimic other conditions with distinct natural courses and treatments [
12]. PNV, distinguished based on type 1 macular neovascularization in eyes exhibiting pachychoroid features, is distinct from neovascular AMD owing to the presence of pachychoroid features and the absence of drusen [
2].
At present, anti-VEGF therapy is the gold standard for neovascular AMD, and its efficacy in addressing PNV has also been documented [
13-
15]. The MINERVA study revealed that eyes with CNV secondary to CSC treated with ranibizumab exhibit superior visual improvements compared with sham-treated eyes, despite some eyes being refractory to intravitreal anti-VEGF monotherapy [
16]. Recently, half-fluence PDT combined with anti-VEGF injections was also administered to patients with PNV [
17]. This combined treatment has shown promise in promoting the regression of macular neovascularization and reducing vascular permeability of the choriocapillaris, thereby influencing choroidal thickness and potentially aiding in the absorption of retinal fluid [
18].
While CSC is typically viewed as a self-limiting condition (acute CSC), it can persist for 6 months or more (chronic CSC), leading to irreversible changes in the retina, RPE, and choroid [
19]. The development of neovascularization in chronic CSC is a significant complication, contributing to reduced visual acuity during long-term follow-up [
20]. The incidence of new-onset PNV in patients with chronic CSC in this study was 11.8%, which is consistent with that reported previously for type 1 CNV in chronic CSC (13.1%-17.6%) [
9,
21-
23].
As a single consensus has not been reached on the treatment protocol for chronic CSC, the initial treatment decision is currently largely reliant on the ophthalmologist’s discretion and the patient’s preference. In a recent large international multicenter study involving 1,719 patients with CSC [
24], the initial treatments varied significantly: anti-VEGF injections in 8.8%, focal laser in 11.8%, PDT in 18.3%, micropulse laser in 2.0%, and conservative or other treatments in 59.1%. These results demonstrate that various treatment options are possible depending on ophthalmologists’ decisions. However, cases of neovascular CSC are more likely to require treatments such as PDT, laser, and anti-VEGF injections to resolve SRF [
25,
26]. While no standard treatment algorithm has been established for PNV, the use of PDT and anti-VEGF agents individually or in combination has been shown to exhibit similar clinical efficacy [
17,
27-
29]. The rationale is that repeated anti-VEGF injections may have a long-term effect on choroidal thickening and congestion [
30]. Nonetheless, Hata et al. [
31] reported that VEGF levels in the aqueous humor of eyes with PNV were lower than those in eyes with AMD, suggesting that anti-VEGF treatment may be less effective for PNV than for AMD. The optimal dosing and timing for treating these conditions remain unclear, necessitating further studies to establish the best approach.
In our study, a significantly higher proportion of patients with initial PNV (80%) received anti-VEGF injections as initial treatment, and treatment initiation tended to be the quickest in this group. Nevertheless, the time to SRF resolution was notably prolonged within this group. These observations align with a recent investigation concerning the effects of anti-VEGF injections in eyes with PNV associated with chronic CSC [
6]. This study indicated that 38.2% of patients with PNV associated with CSC exhibited residual sub/intraretinal fluid at the conclusion of the follow-up period. The persistent presence of SRF in PNV cases can be attributed to the existence of mature vessels within neovascular membranes in pachychoroid disease [
32], which typically exhibit reduced responsiveness to anti-VEGF injections [
33].
In this study, patients with initial PNV also tended to receive more anti-VEGF injections until they achieved initial resolution compared with the number of injections for those without baseline PNV, although this difference was not statistically significant (3.76 ± 5.90 vs. 1.64 ± 2.06 vs. 1.74 ± 4.33 injections to achieve first resolution). These results suggest that in this real-world study, treatment strategies were tailored based on the presence or absence of PNV at baseline. However, the incidence of SRF recurrence was significantly higher in group 2 than that in group 1 (29.2% vs. 64.3%,
p = 0.04). This difference implies the need for a more assertive approach when treating patients with new-onset PNV, who have been less aggressively treated than patients with initial PNV, potentially contributing to a higher recurrence rate in the former group. Our study emphasizes the need for routine screening using OCTA or FAG to detect the emergence of PNV in individuals initially diagnosed with CSC without PNV. With aggressive treatment, PNV is known to be associated with better anatomical outcomes than AMD, with a lower incidence of macular atrophy and fewer intravitreous injections required annually [
15,
34].
Data from previous literature show that CSC complicated with PNV is associated with poor visual acuity and low rates of complete fluid resolution [
9]. A recent multicenter observational study revealed a significantly worse final BCVA in patients with CSC with macular neovascularization and a higher incidence of incomplete fluid resolution, compared with those in the non-neovascular group [
9]. In our study, substantial improvements in BCVA were evident in groups 1 and 3 at the end of follow-up, compared with baseline observations, whereas group 2 did not exhibit significant improvements. Hence, the early identification of neovascularization using OCTA and prompt therapeutic interventions are recommended to mitigate unfavorable outcomes.
Previous studies have identified risk factors for PNV development in patients with CSC, including female sex, systemic hypertension, ellipsoid zone disruption, the presence of a double-layer sign on OCT, and prior PDT [
9,
21,
22]. However, in our study, none of these risk factors demonstrated significant associations with an increased risk of new-onset PNV. The baseline subfoveal choroidal thickness was also similar across the three groups, which is consistent with previous literature indicating that subfoveal choroidal thickness is not a risk factor for the development of neovascularization in patients with CSC [
35]. Future investigations based on larger sample sizes are warranted to obtain more conclusive results.
Our study has three main limitations. First, the retrospective study design and a relatively small sample size limited the generation of statistically significant data. A larger sample size would increase the robustness of the statistical analysis. Second, potential selection bias could have been introduced as this study focused on patients receiving tertiary medical services and those who underwent FAG or OCTA examinations for CSC at their initial evaluation. Third, only patients with diagnostic codes for CSC at their initial clinic visit were included in this study. Consequently, older patients presenting with PNV may have been more frequently misdiagnosed with AMD, potentially contributing to the similar baseline ages observed across the three groups. In contrast, a similar retrospective study reported an older age in patients with CSC who later developed CNV during follow- up, compared with those without neovascularization until the end of follow-up [
35]. Lastly, the introduction of OCTA at our center in 2016, known for its superior sensitivity in detecting choroidal neovascularization compared with FAG/ICG [
36], could have resulted in variability in the diagnostic approach for patients who visited before 2016 and solely underwent FAG/ICG examinations for detecting PNV. Despite these limitations, our study offers valuable real-world insights into the long-term recurrence and visual outcomes of CSC complicated with PNV.
In conclusion, a substantial proportion of CSC cases can progress to PNV. In our study, patients with new-onset PNV were associated with poorer visual outcomes and higher incidences of SRF recurrence than patients with initial PNV or those who did not develop PNV during the follow-up. Our findings underscore the importance of routine screening using OCTA or FAG for the timely and accurate diagnosis of PNV in individuals initially diagnosed with CSC without PNV, aiming to achieve favorable visual and anatomical prognoses.