Comparing the effectiveness of treating branch retinal vein occlusion with a novel one and stepped pro re nata treatment protocol

Abstract

This study aimed to investigate the anatomical and functional outcomes of branch retinal vein occlusion (BRVO) eyes treated with anti-vascular endothelial growth factor (anti-VEGF) injections using a novel one and stepped pro re nata protocol. This retrospective case series evaluated the electronic medical records from 22 BRVO patients who were treated with anti-VEGF agents under our novel “One and Stepped PRN” protocol at a single tertiary medical center between January 2016 and October 2022. Outcomes of interest included best-corrected visual acuity and central retinal thickness. Twenty-two treatment-naive BRVO eyes (14 males, 8 females) were included. The mean age was 65.82 ± 10.88 years. Average follow-up was 54.45 ± 7.65 weeks. 7 (31.81%) received mainly Ranibizumab, and 15 (68.18%) received mainly Aflibercept. The baseline average best corrected distance visual acuity was 45.86 ± 19.46 Early Treatment of Diabetic Retinopathy Study letters, and the baseline average central retinal thickness was 562.5 ± 164.02 μm. The mean number of injections received was 3.54 ± 1.74. Average best corrected distance visual acuity improvement was 23.91 ± 17.36 Early Treatment of Diabetic Retinopathy Study letters (P < .0001) and average central retinal thickness improvement was 245.55 ± 153.31 μm (P < .0001). Our results were comparable to the BRAVO and VIBRANT trials while comparatively using fewer anti-VEGF injections. In summary, our novel anti-VEGF protocol applied under real-world conditions achieved good anatomical and visual outcomes among treatment-naive BRVO eyes.

1. Introduction

Branch retinal vein occlusion (BRVO) is the occlusion of the retinal veins branches due to mechanical compression, vessel wall degeneration, or hypercoagulative status.^[1,2] One of the leading causes of reduction in visual acuity among BRVO patients is macular edema (ME).^[3] The other complications of BRVO can include neovascular development. This can result in vitreous hemorrhage, tractional retinal detachment, or neovascular glaucoma.

Intravitreal injection(IVI) of Anti-vascular endothelial growth factor (anti-VEGF) therapies become the mainstay treatment to manage BRVO eyes after being approved by the Food and Drug Administration of the United States between 2010 and 2014.^[4,5] Several trials have supported the clinical utility of anti-VEGF for BRVO patients. This includes the BRAVO study (monthly injection for 6 months and received injection if needed during month 6–11), the BRIGHTER study (In the Ranibizumab monotherapy arm, at least 3 monthly injections were given until vision was stable. The next retreatment will started only when vision decreased due to disease activity during monthly exam.), and the VIBRANT study (In the IVI group, injections were given every 4 weeks until week 24 and then every 8 weeks until week 48 with rescue grid laser if needed at week 36.^[6–8]) Most of the protocols of these studies started with a 3 to 6 monthly injection as loading phase regardless of edema condition, then proceeded with different follow up protocols. Despite their clear benefits, frequent anti-VEGF therapies may constitute a burden, including a financial burden and the risk of infection, for both patients and clinicians alike.

In order to adjust to real world burden, protocols like Treat & Extend and pro re nata (PRN) developed from treating age-related macular degeneration (AMD) and diabetic macular edema were applied to treat BRVO.^[9,10] However, studies have reported that BRVO patients received no treatment could still achieve good outcomes.^[11,12] This was described in a review that reported improvement in the visual acuity^[12] among those with untreated BRVO eyes.^[13] Such clinical results obtained from untreated BRVO eyes have been hypothesized to be due to the formation of collateral circulation as an adaption to bypass the total capillary bed obstruction seen among BRVO eyes.^[14] Such collaterals can salvage residual retina function as well as decrease the occurrence of edema. This indicates a good prognosis from the perspective of pathophysiology and hemodynamics of BRVO.

Under the finding that certain portions of patients under natural clinical course with good outcomes combined with the evidence of the collateral vessel maturation which is different from AMD and diabetic macular edema, this raise the question of whether we truly need that many injections of IVI-antiVEGF and whether these patients really require monthly follow-up. Unfortunately, available evidence on the different PRN protocols for BRVO eyes has been limited by their variability in study designs and protocols. For example, there was a lack of consensus on the use of a loading dose^[15] as well as the clinical indications for IVI usage.^[16] Also, none of these studies assessed the outcomes of extending the follow-up period instead of monthly intervals.

This study aimed to assess the anatomical and functional outcomes of BRVO eyes with ME treated with our novel “One and Stepped PRN” anti-VEGF IVI protocol. We also sought to compare the efficacy of our novel PRN protocol with results from the BRAVO and VIBRANT trials.

2. Material and methods

2.1. Study population

This retrospective and single institutional study involved the evaluation of electronic medical records of treatment naïve BRVO patients undergoing our proposed “One and Stepped PRN” anti-VEGF protocol. All patients underwent a routine ophthalmic examination at every visit. This included best-corrected visual acuity (BCVA) assessment and evaluation by spectral-domain optical coherence tomography (OCT) (Heidelberg Engineering, Heidelberg, Germany). The study period was from January 2016 to October 2022. This study has been approved by the Medical Center Research Ethics Committee from the China Medical University Hospital (CMUH112-REC1–044). It has been conducted in accordance with the principles of the Declaration of Helsinki. Due to the use of anonymized data obtained during routine clinical care, the informed consent was waived by the Ethics Committee’s approval.

The inclusion criteria included patients of all ages diagnosed with BRVO and followed up for at least a year between January 2016 and October 2022. Furthermore, we only recruited patients with BRVO who received at least one dose of anti-VEGF within the study period.

The exclusion criteria included patients with a history of other ocular diseases, including retina diseases (e.g., diabetic retinopathy and AMD). We also excluded those with a previous history of any anti-VEGF IVI therapies. However, if cataracts developed during the follow-up period, we only excluded them if the patient did not receive cataract surgery before the end of the follow-up time. Also, we only excluded patients with epiretinal membrane located within the parafovea (central 2.5 mm in diameter).

2.2. Treatment regime

Our proposed “One and Stepped PRN” protocol for BRVO patients consisted of 2 phases: (1) a monthly treatment phase and (2) a stepped PRN phase. BRVO patients found to have ME will receive IVI of anti-VEGF during monthly follow up until the edema was resolved. When the ME has been resolved, the patient will be shifted to the stepped PRN phase of the protocol. As part of this modified PRN phase of the protocol, patients will initially return 1 month after the last IVI injection. This is followed by 2 months, then 3 months, until the maximum of 4 months after the initial date was reached (see Fig. S1, Supplemental Digital Content, https://links.lww.com/MD/P693). Unlike treat and extend protocol, IVI was not given during all of these follow up time. However, if there was a reoccurrence of ME, monthly treatment of IVI anti-VEGF will be reinitiated until another resolution of ME is achieved 1 month after the last injection. The next follow-up period would be the longest interval the patient has had previously. For instance, if a patient has extended his follow up period to 3 months and ME occurs after 4 months return. Again his ME is under control after several monthly IVI injections. Then his next follow up period will be 3 months. For our study, ME was defined as central retinal thickness (CRT) >300 μm assessed by OCT or the presence of other evidence of intraretinal fluid or subretinal fluid as previous studies indicated.^[17] Whether patients received combined posterior sub-tenon’s triamcinolone acetonide (PSTA) injection was according to doctor’s clinical judgement.

To analyze the effects of different anti-VEGF IVI therapies, we classified our patients into 3 groups based on their treatment. Group 1 was composed of BRVO patients who only received Bevacizumab (Avastin) injections. Group 2 are patients who either received Ranibizumab (Lucentis) only or received Bevacizumab first and then Ranibizumab. Regarding Group 3, we classified patients who either received Aflibercept (Eylea) only or received Bevacizumab first and then Aflibercept. The reimbursement policy of the Taiwan National Health Insurance towards anti-VEGF IVI usage influenced the organization of our 3 treatment subgroups. To qualify for anti-VEGF IVI reimbursements, patients must first undergo review by the Taiwan National Health Insurance.^[18] Therefore, Taiwanese patients often opt to self-pay for Bevacizumab treatment during this review period as a cost containment option.^[12] Additionally, under Taiwan National Health Insurance, the application criteria for both Ranibizumab and Aflibercept are identical. There is no physician preference in the choice of medication.

2.3. Outcome measures

The primary outcome of interest was the mean change from baseline BCVA recorded via Early Treatment Diabetic Retinopathy Study (ETDRS) letters and CRT changes assessed by OCT.

In terms of baseline demographics, we collected factors such as age, certain medical comorbidities such as hypertension (HTN) and diabetes mellitus (DM) status, initial BCVA, and CRT. Other variables of interest included the total number of IVIs received, the final BCVA, and CRT.

2.4. Statistical analysis

BCVA changes and CRT changes were calculated and compared by paired t test. Subgroup analyses were classified based on the type of IVI and the presence of DM and HTN. Continuous variables were displayed by means and standard deviations (SD). Categorical variables were displayed by portion and percentage. The Kaplan–Meier plot displayed the incidence of recurrence. Results from our study were compared with previous trials (BRAVO^[6] and VIBRANT^[8]). Student t test with unequal variance was conducted for continuous data and chi-square test and Fisher exact test for categorical data. Regression analyses were performed to evaluate the potential risk factors.

All statistical analyses were two-tailed, and statistically significant results were considered when the P-value was less than.05. STATA 1.4.2 was used for statistical analysis.

3. Results

3.1. Baseline data

The demographic data is shown in Table 1. Twenty-two treatment naïve BRVO eyes (36.3% female; average age 65.82 ± 10.8 years) were recruited for this study. Regarding anti-VEGF IVI therapies, none of our patients received Bevacizumab (Group 1) only. Seven patients received Ranibizumab as their primary treatment (and were listed under Group 2). Fifteen patients received Aflibercept as their primary treatment (and were listed under Group 3). The average follow-up duration was 54.45 ± 7.65 weeks. 9 (40.9%) of our patients had DM, and 11 (50%) had HTN at baseline. The average BCVA at baseline was 45.86 ± 19.46 ETDRS letters, and the average initial CRT at baseline was 526.5 ± 164.02 μm. The average final BCVA improvement was 23.91 ± 17.36 ETDRS letters and the average CRT improvement was 245.55 ± 153.31 μm.

Table 1.

Baseline demographics and clinical outcomes.

Demographic variable*
Average age, mean (SD)	65.82 (10.88)
Gender, n(%)
Male	14 (63.6)
Female	8 (36.3)
Diabetes mellitus, n(%)	9 (40.9)
Hypertension, n(%)	11 (50)
Portion receiving PSTA, n(%)	6 (27.3)
Average follow up duration, weeks, mean(SD)	54.45 (7.65)
Average number of IVI injections, n, mean(SD)	3.54 (1.74)
Average number of IVI in first 3 months, n, mean(SD)	2 (0.87)
Average initial BCVA, mean(SD)	45.86 (19.46)
Average final BCVA, mean(SD)	69.77 (29.58)
Average BCVA improvement, mean(SD)	23.91 (17.36), P < .0001
Average initial CRT, μm, mean(SD)	562.50 (164.02)
Average final CRT, μm, mean(SD)	280.95 (53.61)
Average CRT improvement, μm, mean(SD)	245.55 (153.31), P < .0001

^*0 patient in group 1.

BCVA = best-corrected distance visual acuity, display by ETDRS letter score; CRT = central retinal thickness; IVI = intravitreal injection; HTN = hypertension, PSTA = posterior sub-tenon’s triamcinolone acetonide; SD = standard deviation.

3.2. Treatment effects

Among our 22 patients recruited for this study, 7 received Ranibizumab (Group 2) as part of their anti-VEGF therapy, and 15 received Aflibercept (Group 3) as part of their anti-VEGF therapy. None of our patients received Bevacizumab only (Group 1) (see Table 2). No significant statistical difference was found between the patients from Group 2 and Group 3 in terms of initial BCVA (P = .12), initial CRT (P = .93), and the improvement of BCVA (P = .93) and CRT (P = .27) (see Table 2).

Table 2.

Baseline characteristics and treatment effects among classified groups.

	Group 2 (first Bevacizumab and then Ranibizumab or all Ranibizumab) n = 7	Group 3 (first Bevacizumab and then Aflibercept or all Aflibercept) n = 15	P value*
Average amount of IVI	3.86 (1.95)	3.73 (1.58)	P = .87
Average initial BCVA, ETDRS letter score, mean (SD)	55.43 (20.92)	41.4 (17.71)	P = .12
Average BCVA improvement, ETDRS letter score, mean (SD)	18.43 (16.95)	26.47 (17.52)	P = .93
Average initial CRT, μm, mean (SD)	548.29 (231.64)	516.33 (130.41)	P = .68
Average CRT improvement, μm, mean (SD)	262.43 (197.38)	237.67 (135.46)	P = .27

^*Two sample t test.

: BCVA = best-corrected distance visual acuity; CRT = central retina thickness; IVI = intravitreal injection, SD = standard deviations.

The average injections of anti-VEGF given were 3.54 ± 1.74 (2 ± 0.87 in the first 3 months) (see Table 1). The overall average BCVA improvement was 23.91 ± 17.36 ETDRS letters (P < .0001), and the overall average CRT improvement was 245.55 ± 153.31 μm (P < .0001) (see Table 1). Figure 1 shows the Kaplan–Meier survival curve (see Fig. 1) for all eyes recruited. Our Kaplan–Meier curve revealed a median of 12 weeks of stable condition without ME after the first resolution.

Figure 1.

Kaplan–Meier plot of the incidence of recurrence for branch retinal vein occlusion patients who received resolution. Dotted line: median time of stable condition without macular edema after first resolution. (12 weeks).

Patients receiving PSTA had baseline BCVA that was worse than those who didn’t receive PSTA (31.38 ± 6.26 vs 54.14 ± 4.14, P = .005). However, initial CRT (P = .41), BCVA improvement (P = .80) and CRT improvement (P = .69) did not significantly differ between these 2 groups.

3.3. Multivariable regression analysis

Multivariable regression analysis was also performed to determine what factors may influence the improvement of final BCVA and CRT parameters compared to the baseline. We found that the initial baseline BCVA was a statistically significant negative predictor for improvement in BCVA at the final follow-up (P = .03). In other words, patients with worse baseline BCVA reported more improvement in BCVA at follow-up. We also found that the initial CRT was a statistically significant predictor for improvement in CRT at the final follow-up (P < .001). In other words, patients with higher baseline CRT reported more improvements in terms of CRT at the final follow-up (see Table 3). The factors of age, time from symptoms onset, and associated medical comorbidities of DM and HTN were found to have no significant effect on the overall BCVA and CRT improvement from baseline.

Table 3.

Association between clinical factors and improvement.

	Crude association, P value	Adjusted association, P value
Improvement in BCVA
Initial BCVA(ETDRS letters)	.08	.03*
Age	.46	.83
Time from symptoms to treatment	.99	.27
Diabetes mellitus	.18	.15
Hypertension	.51	.45
Improvement in CRT
Initial CRT	<.001*	<.001*
Age	.48	.98
Time from symptoms to treatment	.04*	.15
Diabetes mellitus	.36	.58
Hypertension	.60	.06

^*P value ≦ .05.

BCVA = best corrected visual acuity; CRT = central retinal thickness; ETDRS = Early Treatment Diabetic Retinopathy Study, HTN = hypertension.

3.4. Statistical comparisons with other trials

Several clinical parameters were compared between our results and the 2 previously published anti-VEGF-related trials (the BRAVO and VIBRANT study). The compared parameters included the average amount of IVI used, average initial and improvement in BCVA, and average initial and improvement in CRT.

Compared to the BRAVO study,^[6] the number of anti-VEGF injections used within the first year from our Group 2 (Ranibizumab) participants was significantly lower in number (8.4 vs. 3.86 ± 1.95, no SD in BRAVO for P-value calculation). In terms of other parameters of interest, there was no significant difference in initial BCVA between the BRAVO study and our group 2 participants (53.0 ± 12.5 vs. 55.43 ± 20.92, P = .77), initial CRT (551.7 ± 223.5 vs. 548.29 ± 231.64, P = .97), BCVA improvement (18.3 ± 14.89 vs. 18.43 ± 16.95, P = .98) and CRT improvement (347.4 vs. 262.43, no SD in BRAVO for P-value calculation) (see Table 4).

Table 4.

Baseline characteristics and treatment effects between Group 2 and BRAVO 12 months trial.

	Group 2 (first Bevacizumab and then Ranibizumab or all Ranibizumab) n = 7	BRAVO 12 months trial N = 131	P value*
Average amount of IVI	3.86 (1.95)	8.4 (–)†	–
Average initial BCVA, ETDRS letter score, mean(SD)	55.43 (20.92)	53.0 (12.5)	P = .77
Average BCVA improvement, ETDRS letter score, mean (SD)	18.43 (16.95)	18.3 (14.89)	P = .98
Average initial CRT, μm, mean (SD)	548.29 (231.64)	551.7 (223.5)	P = .97
Average CRT improvement, μm, mean (SD)	262.43 (197.38)	347.4 (–)†	–

^*Two sample t test.

^†Inadequate information to obtain standard deviation.

BCVA = best corrected visual acuity; CRT = central retinal thickness; IVI = intravitreal injection, SD = standard deviations.

Compared to the VIBRANT 52 weeks study,^[8] the number of anti-VEGF injections received was significantly more than our Group 3 (Aflibercept) participants (9.0 ± 1.8 vs. 3.73 ± 1.58, P < .0001). Regarding the proportion of patients with different degree of vision gain or vision loss, we found no significant difference between VIBRANT and our Group 3 (P = .22) (see Tables 5 and 6). Regarding other parameters of interest (e.g., initial BCVA and CRT), there was insufficient baseline data from the VIBRANT trial to calculate any statistical comparisons.

Table 5.

Baseline characteristics and treatment effects between Group 2 and VIBRANT 52weeks trial.

	Group 3 (first Bevacizumab and then Aflibercept or all Aflibercept) n = 15	VIBRANT 52weeks trial N = 91
Average amount of IVI	3.73 (1.58)	9.0 (1.8)	P < .0001*
Portion of patients with improvement < 15ETDRS letters(%)	2 (13.33%)	32 (35.16%)	P = .22†
Portion of patients with improvement ≧ 15 ETDRS letters(%)	12 (80%)	52 (57.14%)
Portion of patients with ETDRS letters loss(%)	1 (6.67%)	7 (7.7%)
Average initial CRT, μm, mean(SD)	516.33 (130.41)	–‡	–‡
Average CRT improvement, μm, mean(SD)	237.67 (135.46)	283.9 (–)‡	–‡

^*Two sample t test.

^†Fisher exact test.

^‡Not available data for further calculation.

CRT = central retinal thickness, IVI = intravitreal injection, SD = standard deviations.

Table 6.

Baseline characteristics and treatment effects between Group 2 and VIBRANT 52weeks trial (detailed version).

	Group 3 (first Bevacizumab and then Aflibercept or all Aflibercept) n = 15	VIBRANT 52weeks trial N = 91
Average amount of IVI	3.73 (1.58)	9.0 (1.8)	P < .0001*
Portion of patients with improvement < 5 ETDRS letters (%)	2	4	P = .27†
Portion of patients with improvement 5–9 ETDRS letters (%)	1	7
Portion of patients with improvement 10–14 ETDRS letters (%)	1	21
Portion of patients with improvement 15–29 ETDRS letters (%)	10	39
Portion of patients with improvement ≧ 30 ETDRS letters (%)	7	13
Portion of patients with ETDRS letters loss < 5ETDRS letters (%)	0	2
Portion of patients with ETDRS letters loss 5~9ETDRS letters (%)	1	2
Portion of patients with ETDRS letters loss 10~14ETDRS letters (%)	0	1
Portion of patients with ETDRS letters loss ≧ 15 ETDRS letters (%)	0	2

^*Two sample t test.

^†Fisher exact test.

IVI = intravitreal injection.

4. Discussion

This was a novel retrospective case series to investigate the effects of our novel “One and Stepped PRN” anti-VEGF protocol over naïve BRVO eyes.

4.1. Novel findings

We found that the One and Stepped PRN anti-VEGF protocol significantly improved average BCVA and average CRT. The results also revealed a median stable condition in terms of no ME after the first resolution after 12 weeks. Furthermore, anatomical and functional parameters assessed within our study did not significantly differ among the different anti-VEGF agents used (i.e., Bevacizumab, Ranibizumab, and Aflibercept). Our multivariable regression analysis also revealed that initial BCVA and initial CRT have significantly affected final BCVA and CRT improvements, respectively.

In terms of PSTA, this was designed as an adjuvant therapy for those with a relatively worse initial condition. This could be seen from our results, as the PSTA group reported worse initial BCVA but had similar gains in improvements (BCVA and CRT) at final follow-up. This indicated that such patients require other supportive treatments to obtain similar clinical improvements. Further studies are needed to devise a criteria for BRVO patients who need additional PSTA.

4.2. Clinical implications

The One and stepped PRN protocol used fewer IVI injections compared to notable studies like BRAVO and VIBRANT. The vessel occlusion in BRVO results in a hypoxic state within the intraocular environment. This triggers the release of several cytokines, including endothelin-1 and VEGF. Such cytokines will promote edematous changes due to increased vascular permeability within the affected intraocular areas. Interestingly, BRVO eyes have been known to adapt to such hypoxic stress through the formation of collateral circulation. However, since VEGF plays an important role in vessel formation, it is still under discussion whether anti-VEGF might theoretically inhibit collateral development. Such impairment on collateral formation by anti-VEGF agents may potentially prolong the duration of the BRVO disease.^[19] One of the main purposes behind our “One and Stepped PRN” protocol was to promote the formation of such collaterals by decreasing the number of anti-VEGF used among BRVO eyes. The other implications of our protocol were that it could also potentially reduce patient burden and improve capacity among ophthalmic clinics.^[20] Examples of treatment burdens include demand on resources, annual financial cost(252,000 dollar in BRAVO trials, 270,000 dollar in VIBRANT trials, 106,200 dollar in our protocol, for each IVI cost 3000 New Taiwan Dollar according to self-paid price in Taiwan.), travel time by patients, risk of infection during procedure.

4.3. Comparison to other studies

Our results showed that patients possessing poorer baseline BCVA were significantly associated with greater gains in BCVA at final follow-up after anti-VEGF treatment. Furthermore, a thicker CRT at baseline was significantly associated with greater CRT improvement at the final follow-up after anti-VEGF treatment. These results were similar to previous published studies.^[6,7,21] We deduced that participants with thicker CRT and worse BCVA at baseline would have greater opportunities for improvement and are more likely to show improvement in certain anatomical and functional parameters.

While age and HTN were associated with a higher risk of developing BRVO, according to the Eye Disease Case-control Study Group study, our findings indicated that such risks did not translate into actual statistical effects on BCVA gains or posttreatment CRT reductions. There are factors that may have influenced our results such as how patient included in the study controlled their medical comorbidities.^[3] These form interesting avenues for future study to further explore.

The BRAVO trial investigated clinical effects of ranibizumab on patients with ME after BRVO by 6 monthly injections of ranibizumab. The average BCVA improvement from BRAVO regarding the ETDRS letter score was 18.3. The average CRT improvement from the 12th month was 347.4 μm. Our results have been comparable to BRAVO regarding average BCVA and CRT improvement. This has clinical implications when considering that the mean amount of IVI used in BRAVO trail is 8.4 injections but for our Group 2 (Ranibizumab) was 3.86.

Additionally, it should be noted that our study’s CRT improvement seemed less in magnitude compared to the BRAVO trial. However, this hints towards a ceiling effect in terms of CRT improvement as the natural thickness of the fovea has been estimated to be around 230 to 240 μm.^[22] Therefore, it can be hypothesized that eyes with thicker CRT would have greater room for CRT improvement. However, differences in study design between ours and BRAVO could make direct comparisons difficult. One example is concerning the difference in treatment threshold for CRT. For example, patients in BRAVO received IVI Ranibizumab only if CRT was >250 μm. Comparatively, our treatment threshold was at 300 μm.

In terms of the VIBRANT trial, their results showed that at 52 weeks, 57.14% of their patients had improvement in BCVA with an average amount of 9 intravitreal injections. This was interesting as Group 3 (Aflibercept) in our study achieved similar improvement in BCVA while using fewer IVI than the VIBRANT trial (3.73 vs 9). Regarding average CRT improvement, VIBRANT achieved a greater reduction in average CRT compared to ours (283.9 vs 237.67).

There have also been other studies that have supported the application of a PRN protocol without a loading dose in the initial stages of BRVO treatment.^[23–25] In one study, Chen et al compared BRVO patients receiving Conbercept under 1 + PRN protocol and 3 + protocol.^[23] They reported that there was no significant difference between the 2 groups in terms of BCVA and CRT improvement. In another prospective randomized study by Tang et al, the results again validated adopting a treatment protocol that reduces the number of IVI Ranibizumab needed to treat ME.^[24] Our One and Stepped PRN protocol further extend follow up period providing no ME occur. This is still different from treat and extend protocol as no need of IVI during our extending PRN follow up period. Another interesting aspect was that our study reported a median stable condition without ME after the first resolution was around 12 weeks. The results were similar to those of Noma et al, who reported the mean period between the first and second IVI aflibercept was 92.5 ± 40.8 days.^[26]

5. Strength and limitation

This is the first real-world data from an Asian country that assessed the outcomes from using a modified PRN protocol to treat naïve BRVO eyes. Furthermore, we also included a multi-regression analysis of predictors that involved anatomical and functional parameters. However, several limitations of our study deserve some discussion. These included no randomized control group, retrospective nature, limited follow-up, inclusion of patients with PSTA treatments, and small sample size. Other limitations include the lack of analysis for laboratory findings such as HbA1c levels, total cholesterol, and other related lipid serum levels. In addition, we did not evaluate OCT parameters such as the integrity of the ellipsoid zone, the inner retinal layers and effect of macular perfusion status may have on our results. Although most of the protocols in previous trials did not assign IVI according to the presence of these parameters or not. These parameters do affect the prognosis and treatment response^[27] and is worth for further research.

6. Conclusion

In conclusion, we have shown that treatment naïve BRVO eyes treated with One and Stepped PRN protocol significantly improved BCVA and CRT. We also showed that having a worse initial BCVA and CRT predicted further improvements among these functional and anatomical parameters. Furthermore, there was no significant difference in BCVA and CRT improvements between our study and the BRAVO and VIBRANT trials.

Author contributions

Conceptualization: Po-Yu (Jay) Chen, Chun-Ting Lai, Chun-Ju Lin, Ning-Yi Hsia, Wen-Lu Chen, Peng-Tai Tien, Jane-Ming Lin, San-Ni Chen, Yi-Yu Tsai.

Data curation: Po-Yu (Jay) Chen, Alan Y. Hsu, Chun-Ting Lai, Chun-Ju Lin.

Formal analysis: Po-Yu (Jay) Chen.

Investigation: Po-Yu (Jay) Chen, Alan Y. Hsu, Chun-Ting Lai, Ning-Yi Hsia, Wen-Lu Chen, Peng-Tai Tien, Jane-Ming Lin, San-Ni Chen, Yi-Yu Tsai.

Methodology: Po-Yu (Jay) Chen.

Project administration: Chun-Ting Lai.

Software: Po-Yu (Jay) Chen.

Supervision: Chun-Ting Lai, Wen-Lu Chen, Yi-Yu Tsai.

Writing – original draft: Po-Yu (Jay) Chen, Alan Y. Hsu.

Writing – review & editing: Alan Y. Hsu, Chun-Ting Lai, Chun-Ju Lin, Ning-Yi Hsia, Wen-Lu Chen, Peng-Tai Tien, Jane-Ming Lin, San-Ni Chen, Yi-Yu Tsai.