The study included 19 eyes from 19 enrolled patients (5 males, 14 females). The mean (± standard deviation) patient age was 62.47 (± 13.48) years, and the mean (± standard deviation) follow-up period was 4.81 ± 2.71 years. During the follow-up period, a total number of 126 anti-VEGF injections were administered (bevacizumab; 119, ranibizumab; 7).
Patients’ detailed diagnoses and received treatment are summarized in Fig. 1. Among the patients enrolled in the study, 5 patients were diagnosed with accompanying CNV and all patients were treated with anti-VEGF injection. Among the 14 patients without CNV, 11 patients received an anti-VEGF injection treatment and 3 patients also received PDT treatment. The remaining 2 CNV-free patients were followed up without any treatment. Among the 14 CNV-free patients, there was a total of 243 inter-visits, 88 (bevacizumab; 81, ranibizumab; 7) were administered intravitreal anti-VEGF injection, and 155 were observed without treatment.
Complete Subretinal Fluid Resolution
Whether complete SRF resolution was obtained at the last visit according to the presence or absence of CNV and treatment was analyzed (Table 1). In 2 CNV-free patients without any treatment, one patient achieved spontaneous complete SRF resolution at the last visit. Among the patients treated, complete resolution was observed in 4 patients in the CNV group (4/5, 80.0%) and in only 2 patients in the CNV-free group (2/12, 16.7%), with the difference being statistically significant (P = 0.028, Fisher’s exact test).
Table 1
Number of patients who achieved complete SRF resolution at the last visit
| Complete SRF resolution | SRF remained | p-value |
CNV free, untreated (n = 2) | 1 (50.0%) | 1 (50.0%) | |
CNV free, treated (n = 12) | 2 (16.7%) | 10 (83.3%) | 0.028* |
CNV, treated (n = 5) | 4 (80.0%) | 1 (20.0%) |
Values are presented as number (%). |
Fisher’s exact test |
*Statistically significant |
SRF, subretinal fluid; CNV, choroidal neovascularization |
Visual Prognosis
In the CNV-free group, visual acuities were improved in patients with complete resolution, while visual acuities showed decreased tendency in the patients without complete resolution (mean ± standard deviation, LogMAR, 0.48 ± 0.35 to 0.62 ± 0.28; P = 0.092, Wilcoxon signed-rank test) (Fig. 2A).
In the CNV group, despite treatment efforts, visual acuities decreased at the final visit compared to those at the first visit, showing marginal significance (mean ± standard deviation, LogMAR, 0.32 ± 0.13 to 0.67 ± 0.27; P = 0.068, Wilcoxon signed-rank test) (Fig. 2B).
Effectiveness Of Anti-vegf Treatment
The effect of anti-VEGF treatment was determined in the inter-visit changes of foveal SRF height between the cases of anti-VEGF versus observation (no treatment).
When the inter-visit change patterns of foveal SRF were divided into increase, stable (within ± 20%), and decrease, patterns were evaluated as 47 (30.3%), 67 (43.2%), 41 (26.5%) cases in cases of observation, and 19 (21.6%), 33 (37.5%), 36 (40.9%) cases in cases of anti-VEGF administration, respectively, with no significant differences (P = 0.243, chi-squared test) (Table 2).
Table 2
Number of inter-visit changes in foveal SRF categorized as increase, stable, and decrease in patients with CNV-free inferior staphyloma
CNV free (n = 14) | More than 20% increase | Stable | More than 20% decrease | Total | P-value |
Observation | 47 (30.3%) | 67 (43.2%) | 41 (26.5%) | 155 | 0.243 |
Anti-VEGF injection | 19 (21.6%) | 33 (37.5%) | 36 (40.9%) | 88 |
Values are presented as number (%). |
chi-squared test |
SRF, subretinal fluid; CNV, choroidal neovascularization; VEGF, vascular endothelial growth factor |
After correction for confounding factors such as the foveal SRF height at the former visit and inter-visit duration, there was no difference in inter-visit SRF change between the observation cases and anti-VEGF cases for all variable types (Fig. 3, Tables 3 and 4). The estimated mean (95% confidence interval) of the absolute value of inter-visit SRF change was + 0.252 (-43.864,44.368) µm in the observation cases and − 11.845 (-71.003,47.314) µm in the anti-VEGF cases, and there was no significant difference between the two groups (P = 0.733, linear-mixed model with unstructured covariance matrix) (Fig. 3, Table 3). The estimated mean (95% confidence interval) of rate of inter-visit SRF change was + 13.738 (-73.363,100.839) % in the observation cases and + 6.270(-111.628,124.167) % in the anti-VEGF cases, and there was no significant difference between the two groups (P = 0.916, linear-mixed model with unstructured covariance matrix) (Table 3). Adjusted proportion of increase, stable (within ± 20%), and decrease cases of inter-visit SRF were 46.0%, 40.5%, and 13.3% in the observation cases and 47.2%, 33.6%, and 19.3% in the anti-VEGF injection cases, respectively, which was statistically not significant between the two groups (P = 0.277, generalized linear-mixed model) (Table 4).
Table 3
Comparison of absolute value and rate of inter-visit foveal SRF changes between observation and anti-VEGF cases in CNV-free inferior staphyloma patients
Absolute value change (µm) |
Variable | Observation (no treatment) Estimated mean (95% CI) | Anti-VEGF injection Estimated mean (95% CI) | p-value |
Unadjusted | + 4.494(-81.685,90.673) | -15.738(-130.546,99.070) | 0.766 |
Adjusted SRF | + 1.072(-42.869,45.012) | -12.597(-71.197,46.002) | 0.696 |
Adjusted duration | + 4.072(-82.521,90.665) | -15.308(-131.097,100.480) | 0.778 |
Adjusted SRF, duration | + 0.252(-43.864,44.368) | -11.845(-71.003,47.314) | 0.733 |
Rate of change (%) |
Variable | Observation (no treatment) Estimated mean (95% CI) | Anti-VEGF injection Estimated mean (95% CI) | P-value |
Unadjusted | + 15.574(-71.078,102.226) | -0.336(-115.315,114.642) | 0.816 |
Adjusted SRF | + 13.661(-73.227,100.548) | + 6.279(-110.669,123.228) | 0.916 |
Adjusted duration | + 15.337(-71.720,102.394) | -0.194(-116.154,115.766) | 0.822 |
Adjusted SRF, duration | + 13.738(-73.363,100.839) | + 6.270(-111.628,124.167) | 0.916 |
Linear mixed model with unstructured covariance matrix |
SRF, subretinal fluid; CNV, choroidal neovascularization; VEGF, vascular endothelial growth factor; CI, confidence interval |
Table 4
Comparison of proportion categorized as ± 20% of inter-visit foveal SRF changes between observation and anti-VEGF cases in CNV-free inferior staphyloma patients
Variable | Observation (no treatment) Estimated proportion (SE) | Anti-VEGF injection Estimated proportion (SE) | p-value |
Unadjusted | | | |
Increase | 0.303 (0.037) | 0.216 (0.044) | 0.062 |
Stable | 0.431 (0.043) | 0.375 (0.054) |
Decrease | 0.265 (0.035) | 0.409 (0.052) |
Adjusted SRF | | | |
Increase | 0.525 (0.069) | 0.508 (0.098) | 0.290 |
Stable | 0.344 (0.062) | 0.300 (0.078) |
Decrease | 0.133 (0.038) | 0.191 (0.061) |
Adjusted duration | | | |
Increase | 0.259 (0.044) | 0.198 (0.043) | 0.073 |
Stable | 0.486 (0.060) | 0.403 (0.062) |
Decrease | 0.259 (0.042) | 0.403 (0.055) |
Adjusted SRF, duration | | | |
Increase | 0.460 (0.078) | 0.472 (0.100) | 0.277 |
Stable | 0.405 (0.078) | 0.336 (0.088) |
Decrease | 0.133 (0.041) | 0.193 (0.062) |
Values are presented as proportion. |
Generalized linear mixed model |
SRF, subretinal fluid; CNV, choroidal neovascularization; VEGF, vascular endothelial growth factor; SE, standard error |
As a result, the anti-VEGF injection did not significantly reduce SRF compared to observation (no treatment). In fact, in many cases, the SRF was increased despite anti-VEGF injection, and the SRF was decreased despite just observation. A representative case is displayed in Fig. 4.
Effectiveness Of Pdt
A total of 3 patients underwent half-dose PDT, and all patients underwent a single PDT session. After the PDT treatment, in three consecutive follow-up visits at 3-month intervals, complete resolution could not be reached at any time, and the decreased SRF was not maintained (Supplementary Fig. 1).