Efficacy and safety of post-scleral injection of triamcinolone acetonide for macular edema in branch retinal vein occlusion: 3 month results from a prospective randomized, controlled and open label clinical trial

Background: To assess the safety and efficacy of post-scleral injection of triamcinolone acetonide (TA) in macular region to treat macular edema secondary to branch retinal vein occlusion (BRVO), and compare it with intravitreal injection of TA and Ranibizumab. Methods: This is a prospective, double masked and single-center study. Thirty-six eyes, one from each participant who was diagnosed with macular edema secondary to BRVO between July 2011 and December 2014, were randomised into three groups to receive post-sclera injection of 20mg/0.05ml TA (PS-TA), intravitreal injection of 2mg TA (IV-TA) and intravitreal injection of 0.5mg Ranibuzmab (IV-Ranibizumab), respectively. Changes of retinal thickness in optical coherence tomography (OCT) in foveal (zone 1) and average macular area (1-9 zones) at month 1 and 3 from baseline were the primary outcomes. The secondary study outcomes include changes of best corrected visual acuity (BCVA) and intraocular pressure (IOP) at month 1 and 3 from baseline and profile of adverse events. Results: The mean thickness changes from the baseline in macular fovea (zone 1) were statistically significant at month 1 in IV-TA and PS-TA groups, respectively (P-value < 0.0005 and P-value = 0.032), but not in Ranibizumab group (P-value = 0.083). At month 3, the mean thickness change from the baseline in macular fovea (zone 1) was statistically significant only in IV-TA group (P-value = 0.001). There was no significant difference in the mean thickness changes in macular region (zone 1-9) from baseline to month 1 or 3 in any group. No statistically significant differences were detected across the three treatment groups in mean thickness changes in macular fovea (zone 1) nor macular region (zone 1-9), over 1 and 3 months from baseline (P-values > 0.05). No Serious Adverse Event was detected in any treatment group during the study period. choroid, Bruch membrane, and retinal pigment epithelium. Experimental results

reduce the incidence of intraocular hypertension significantly, and thus has the advantages of safety compared to IVTA.

Background
Retinal vein occlusion (RVO) is the second most common cause of retinal vascular abnormality after diabetic retinopathy [1]. Eyes suffering from this disease are vulnerable to vision loss and may become blind due to its complications. According to where the occlusion is located, RVO is classified into central retinal vein occlusion (CRVO), hemiretinal vein occlusion (HRVO), and branch retinal vein occlusion (BRVO). It has been reported that the prevalence of BRVO is 0.6-1.6% in group of age 65 to 74 years old, while the incidence is about 2.14/1000 per year in the population over 40 years old [2]. BRVO often occur at arteriovenous crossings, and the artery is anterior to the vein at the occlusion in most cases. The degree of macula involvement determines the level of vision loss, which may be caused by ischemic injury and/or secondary macular edema. Macular edema is commonly seen in BRVO patients with 5-15% incidence rate within the first year of the disease course and is one of the most significant causes of central vision loss in these patients [3,4].
The first line treatment of macular oedema is intravitreal injection of anti-vascular endothelial growth factor (VEGF) [5,6]; however, it may increase the risks of stroke and intraocular infection [7]. In addition, the cost of intravitreal anti-VEGF such as Ranibizumab may create burdens for the patients and social health system due to its high price [8][9][10], particularly in developing countries. Triamcinolone acetonide (TA) is an approved effective treatment for macular oedema secondary to BRVO. However, TA has high profile in intraocular hypertension and cataract formation [11][12][13][14]. These complications are does and time dependant [15]. To minumise the adverse events of elevation of intraocular pressure and cataract formation, we developed the patent set of post-sclera injection for TA delivery (patent number ZL 2013 2 0740202.0). This treatment has direct effect to macular edema. The instrument allows accurate post-scleral injection with small dosage of TA and hydrotropic substance and solutes carrying positive charge would penetrate the sclera more easily [16,17]. This randomized controlled clinical trial is aimed to assess the safety and efficacy of post-sclera injection of TA in macular region in treatment of BRVO with macular edema, comparing with conventional intravitreal injections of TA and Ranibizumab. Exclusion criteria were: uncontrolled IOP of more than 25 mmHg; significant corneal or lenticular opacities, which would affect the assessment of treatment effect and vision improvement; previous intraocular surgery history; the patients take other investigational drugs; intravitreal anti-VEGF treatment within 6 weeks prior to screen visit; the participants were treated with corticosteroid hormone systemically or locally for more than 30 days continiouly within 6 months at the screening visit; the participants were treated with corticosteroid hormone in or around the eyes within 3 months at the screening visit; the particpants received full-retina laser photocoagulation 3 months prior to the screen; myocardial infarction, transient ischemic attack, or cerebrovascular accident within 90 days prior to screen visit; and current pregnancy or lactation. All ocular exclusion criteria refer to the study eye only.

Randomized grouping and therapeutic methods
Participants who were eligible to the study were uniquely numbered from 1-36 according to the sequence of enrollment and randomized into three groups to receive one of the During the procedure, the participants received a topical anesthesia (benoxinate hydrochloride 0.4%) and followed by a topical antibiotic drops (chloramphenicol solution 0.5%). The needle was connected to an ordinary syringe and was inserted through a small incision at the temporal conjunctiva 6 mm from the corneal limbus. The needle was slided in on the surface of the sclera with the tip reaching macular region. Post-sclera injection of 20 mg in 0.5 ml TA was performed. Daily B-ultrasonography was conducted in the first 6 days after the injection to observe drug residue posteriorly to the macula.
Injection of IV-TA and IV-Ranibizumab were performed as an outpatient procedure in procedure rooms with strict sterile conditions. After topical anesthesia with 0.4% benoxinate hydrochloride eye drops, the ocular surface and the lid were disinfected with povidone-iodine. Intravitreal injection was carried out using a 30-gauge needle at 3.5 mm posterior to the limbus (TA 2 mg/0.05 ml or Ranibizumab 0.3 mg/0.05 ml). The injection site was compressed by cotton swab to avoid reflux, and the fundus was examined to rule out any complication and to check perfusion of the retinal artery.

Outcomes assessment
The primary efficacy outcome was the changes of retinal thickness in OCT (Topcon, 3D, Japan) in foveal area (zone 1) and macular area (1-9 zones) at 1 and 3 month compared to baseline. The secondary study outcomes include changes of BCVA and IOP (Cannon, TX-F, Japan) at 1 and 3 month from baseline and the profile of adverse events.
Participant's retina of study eye was scanned along 8 meridian lines with a scan length of 6 mm. The distance (thickness) between the inner surface of the retina at the macular fovea and the light band of retinal pigment epithelium and choroid blood capillary was measured. The average thickness of 1-9 zones is the mean value of the thickness of the 9 zones in macular area. The 9 zones are centered at the macular fovea and numbered from 1-9. 1 represents the macular fovea (zone 1 with a diameter of 1000 µm); 2-5 represent the inner ring (diameter: 1000-3000 µm), including 4 equal zones: temporal side, upper side, nasal side, and lower side; 6-9 represent the outer ring (diameter: 3000-6000 µm), including 4 equal zones: temporal side, upper side, nasal side, and lower side (see Fig. 1A). The secondary efficacy observation outcome included BCVA, which was examined using ETDRS visual acuity chart (Precision Vision, 2425E, USA). The safety outcome include profile of advese event, serious adverse events and IOP changes. The IOP was measured using non-contact tonometer(Cannon, TX-F, Japan) in each visit.

Statistical method
For each of three treatments considered in this study, therapeutic effect on thickness in macular fovea (zone 1; also known as CMT), as well as on average thickness of whole macular area (zone 1-9) at month 1 and 3 separately, was evaluated by comparing the thickness before and after applying the treatment using paired t test. Analysis of Variance (ANOVA) when appropriate, or its Welch alternative for unequal variances [18], was used for comparing change, over 1 and 3 months respectively, in thickness in macular fovea (zone 1), average thickness in macular area (1-9 zones), BCVA and IOP across the three treatments studied. Treatment effects were further investigated using generalised estimating equations [19] that accounted for possible correlation in repeated measurements over time while adjusting for other factors or potential confounders in the study.
Analyses presented here were performed using SPSS statistical software (version 23; IBM Corp, Armonk, NY, USA). A P-value of less than 0.05 was considered statistically significant (ie, using 5% significance level) in all analyses presented in this paper.  Table 1. Furthermore, the IOPs of all the patients were within normal value range of 12.4mmHg to 19.7mmHg.

Intragroup difference of thickness changes
The mean thickness changes (ie, reduction) at month 1 from baseline in macular fovea (zone 1) and macular area (1-9 zones) were 78.6 and 20.7 for PS-TA, 236.3 and 45.3 for IV-TA and 145.8 and 30.0 for ranibizumab as shown in Table 2. Such reduction in CMT (zone 1) was only considered significant for patients treated with PS-TA (P-value = 0.032) and IV-TA (P-value < 0.0005), but not for those treated with ranibizumab (P-value = 0.083). (Fig. 2).

Intergroup difference of thickness changes in macular fovea (zone 1) and average thickness changes in macular area (1-9 zones)
As shown in Table 3, IV-TA group appeared to have the most significant reduction in CMT and mean CMT at month 1 compared to other treatment groups, however, the changes in both CMT and mean CMT across the three treatments were not statistically significant (Pvalues = 0.133 and 0.730, respectively). No significant difference was detected neither in mean change in CMT (zone 1) or mean CMT (zone1-9) at month 3 from the baseline in three groups (P-values = 0.629 and 0.941, respectively). Furthermore, CMT generaly bounced back noticeably from month 1 to month 3 although not statistically significant, while mean CMT did not change from month 1 to 3. (Fig. 3).
The treatemt effect on CMT was further investigated using more sophiscated statistical analysis via GEE method, accounting for possible correlation among repeated measurements over time (month 0, 1 and 3) and also other possible confounding factors, such as age and gender. The results from the analysis are similar to those from ANOVA described above, confirming CMT decreased from the baseline to month 1 with not significant further change from month 1 to month 3, and there was no significant treatment difference on CMT even after adjusted for possible confounding effects from age, gender and baseline BCVA. Similar analysis was also carried out for mean CMT, confirming no significant difference in mean CMT across the three treatments suggested by the ANOVA earlier. However, a significant difference in mean CMT over time (P-value = 0. 022) was detedcted after accounting for correlation among repeated measurements and adjusting for gender, age and baseline BCVA, unlike the findings on mean CMT from the ANOVA. The mean CMT was reduced from the baseline to month 1 with no further change to month 3.

BCVA changes after injection
For each treatment group, average change in BCVA was examined. As shown in Table 4, no significant change was detected either at month 1 or month 3 for PS-TA group. BCVA was improved by 15.4 letters on average over the first month after the treatment for IV-TA group (P-value < 0.0005), however, further improvement from month 1 to 3 was marginal (P-value = 0.058). Similar but slightly less improvement in BCVA was also observed in IV-Ranibizumab group at month 1 and month 3 (P-value=0.034). (Table.5

)
Changes in BCVA at month 1 and 3 were also compared across the three treatments respectively. There was no treatment difference in the changes even after accounting for patients' baseline factors (P-values > 0.05). However BCVA change over time was confirmed by the GEE analysis (P-value < 0.0005), and baseline BCVA was associated with BCVA change over time among the three groups (P-value=0.001).

TA Pharmacokinetics in PS-TA group
TA was seen in post-sclera closeing to macular area in the first 5 days after the injection, and the residue amount gradually diminished with time. On the 6 th day after the injection, no sound shadow was observed posteriorly to the wall of eyeball, indicating the drug had been absorbed (see Fig. 1B).

Advense events (AEs)
The IOPs in three groups were within normal value range at month 1 post-injecton (IV-TA:

Discussion
Macular edema caused by BRVO is a major cause of vision loss, and there is no effective oral medication to treat for the time being [20]. Currently, the first line treatment of macular oedema secondary to BCVO is intravitreal injection anti-VEGF, Howevr, it has some limitations and disadvantages: intravitreal injection of anti-VEGF has the risk of intraocular infection [21] and also creates burdens to patients due to its costs and frequent visit to clinics, especially in developing countries. Tramcinolone has been confirmed to be effective for the treatment of BRVO with macular edema because of its potent antipermeability and anti-inflammatory properties [15,22]. However, complications resulting from intravitreal injection of TA therapy include cataract formation (24.2%), elevation of intraocular pressure (20%), infectious endophthalmitis, noninfectious endophthalmitis, and retinal detachment [23]. Retrobulbar injection or post-Tenon's capsule injection TA has the risk of piercing the eyeball [24]; peribulbar injection of TA yields poor outcomes. This study explores a safe, economic, and effective treatment of BRVO with macular edema by a prospective randomized, controlled clinical trial between conventional treatment and our new technique.
We use a patented method of post-scleral injection of TA to treat macular edema secondary to BRVO. The instrument and method for post-sclera injection described in this study are different from the instrument and method for post-Tenon's capsule injection. In the latter, drugs are injected into the retrobulbar Tenon's capsule, which is relatively far from macular area, therefore the efficacy is poor. However if the injection is too close to the macula, the risks of piercing the eyeball and damage inferior rectus muscle would be increased. The outer surface of the sclera is covered by the loose episcleral fascia, which mainly consists of collagenous fiber embedded with glycosaminoglycan substrate, and the post-sclera collagenous fiber is looser than the ante-sclera part [25]. Therefore, peri-postscleral injection is beneficial to drug accumulation and absorption. To reach the retinal neurepithelium layer, a drug has to panetrate through several membranes including sclera, choroid, Bruch membrane, and retinal pigment epithelium. Experimental results show that the largest molecular weight of substances, to which the sclera is permeable, is 150KDa [26]. TA, whose molecular weight is 434.5 Da, is very easy to pass through it. In our study, a daily B-ultrasonography was performed continuously to monitor TA diminish after post-scleral injection. TA was seen posteriorly to the macula in the first 5 days after the injection, and the residue amount gradually diminished with time. On the 6th day after the injection, no sound shadow was observed posteriorly to the wall of eyeball, indicating the drug had been absorbed fully and TA could exist behind the sclera in macular area for 5 days.
The peri-post-scleral injection is easy to learn and operate. Because vitreous cavity is not involved in this procedure, the risks of intraocular infection could be avoid. Further more as the needle is blunt, the risk of eyeball punctuation is reduced. Inserting the needle through the small incision with minimal trauma at the conjunctiva above the temporal can avoid damage extraocular muscle. In addition, because drugs are directly injected into the post-macula region, the episcleral Tenon's capsule can retain local drug concentration and yield direct effect in curing macular edema. This method also avoids significant AE of IOP elevation that IV-TA usually casued. The study indicated that all the study eyes in PS-TA group had no significant changes of IOP post treatment.
TA is cost-effective and easier to obtain in developing countried such as China.
Furthermore, TA can also stabilize vascular endothelial barrier by reducing the release of prostaglandin through arachidonic acid pathway [27]. It reduce inflammation by suppressing inflammatory cell proliferation and migration and by decreasing the synthesis and release of proinflammatory molecules (prostaglandins, leukotrienes, vascular endothelial growth factor, and intercellular adhesion molecule 1) [28].
This study found that the differences of thickness changes in macular fovea (zone 1) and mean thickness change in macular area (1-9 zones) between PS-TA group and IV- Figure 1 Diagram of macular zones for BRVO patients and B-ultrasonography image after PS-TA A: Diagram of macular zones; the 9 zones shown in the figure are centered at the macular fovea and numbered from 1-9. 1 represents the macular fovea (zone 1 with a diameter of 1000μm); 2-5 represent the inner ring (diameter: 1000-3000μm), including 4 equal zones: temporal side, upper side, nasal side, and lower side; 6-9 represent the outer ring (diameter: 3000-6000μm), including 4 equal zones: temporal side, upper side, nasal side, and lower side. B: OCT for macular area can generate 9 values corresponding respectively to the 9 zones. The OCT scanner scanned along the horizontal meridian lines of optic nerves with a scan length of 6mm. Each region corresponding numerical values; C: Bultrasonography image obtained 5 days after PS-TA treatment; the arrow points to the drug residue posteriorly to the sclera in macular area, and sound shadow can be seen in the rear.