Main

Retinal vein occlusions belong to the most common retinal disorders affecting the macula and reducing central visual acuity.1, 2, 3 In a recent population-based study, retinal vein occlusions were detected in about 0.7% of eyes of adult Chinese aged 40+ years. Branch retinal vein occlusions were about 12 times more common than central retinal vein occlusions (CRVOs), and the non-ischaemic type was about 9 times more common than the ischaemic type. From a pathogenic point of view, a decreased tissue perfusion and an increased hydrostatic pressure within the involved segments as a consequence of the vascular obstruction may lead to intraretinal haemorrhages, exudation of fluid, varying levels of tissue ischaemia, and eventually to intraocular neovascularization, if retinal ischaemia is pronounced.4

Although the Central Vein Occlusion Study Group has shown the beneficial of panretinal laser coagulation for the treatment of neovascularization, there have not been clear therapeutic recommendations for the treatment of macular oedema caused by CRVO.5 In the past 8 years, a change in a paradigm has taken place, now to consider the vitreous cavity as drug reservoir for the treatment of retinal disorders, such as diabetic retinopathy and retinal vein occlusions.6 The first drug, which was intravitreally injected was triamcinolone,7, 8, 9, 10, 11, 12, 13, 14 followed by ranibizumab and bevacizumab.15, 16, 17, 18, 19, 20, 21, 22 Both drugs differ in the spectrum of side effect and potentially in the magnitude and duration of their effect. As it has been unknown so far, which of the drugs may be preferable in which situation, we conducted a retrospective analysis comparing patients with non-ischaemic CRVO with respect to the change in visual acuity and intraocular pressure.

Materials and methods

The clinical interventional comparative retrospective non-randomized study included 72 eyes (72 patients) with non-ischaemic CRVO who consecutively underwent intravitreal injection of either triamcinolone (4.0 mg; 42 patients) or bevacizumab (1.25 mg; 30 patients) between May 2004 and October 2007, and who were followed for at least 3 months after the intravitreal injection. The diagnosis was substantiated by fluorescein angiography and optical coherence tomography showing significant cystoid macular oedema without marked retinal ischaemia, as defined by the Central Retinal Vein Occlusion Study Group.5 It was the decision of the attending retinologist whether triamcinolone or bevacizumab was injected, with the same retinologist using the same drug during the whole study period. The mean age was 55.64±16.26 years (mean±SD) in the triamcinolone group, and it was 54.67±15.50 years in the bevacizumab group without a significant difference between the two study groups (P=0.80) (Table 1). In a similar manner, both group did not vary significantly in the self-reported duration of the symptoms (P=0.51), and visual acuity (P=0.63), intraocular pressure (P=0.37), macular thickness (P=0.71) at baseline, and the number of patients with peripheral retinal laser coagulation therapies (P=0.62) (Table 1).

Table 1 Baseline parameters (mean±SD) of patients with non-ischaemic central retinal vein occlusions and treated by an intravitreal injection of triamcinolone (4 mg) or bevacizumab (1.25 mg)
Full size table

Inclusion criteria were significant macular oedema as measured by optical coherence tomography, loss of visual acuity, and macular vessel leakage in fluorescence angiography. Exclusion criteria were signs of any other fundus diseases (such as diabetic retinopathy); signs of non-perfusion or ischaemia, defined as neovascularization on the disc or elsewhere, iris neovascularization, or more than 10 disc areas of retinal non-perfusion detected by fluorescein angiography; and any earlier treatment (except of retinal laser coagulation) of the retinal vein occlusions, such as haemodiluting therapy or intravitreal injection of steroids or other antiangiogenic or antioedematous drugs. The study was approved by the local Institutional Review Board, and informed consent was obtained from every patient. The off-label use of bevacizumab and its potential risks and benefits were discussed in detail with the patients.

The technique of the intravitreal injections was similar as already reported in detail previously.7 The preservatives were removed. We applied a dosage of 4 mg of triamcinolone or of 1.25 mg of bevacizumab. At baseline, all patients underwent an ophthalmological examination including refractometry with assessment of best-corrected visual acuity, applanation tonometry, ophthalmoscopy, fluorescein angiography, and optical coherence tomography for measurement of the macular thickness.

After the intravitreal injection, the patients were scheduled to be re-examined at 1 day, 3 days, 1 month, 2, 3, 6 months, and 1 year after the injection. If postoperative problems or complications occurred, the follow-up examinations were carried out in shorter intervals. If during the follow-up the patients underwent ocular surgeries, only the results of the examination carried out earlier to that surgery were taken for the statistical analysis. The patients received re-injections when the macular oedema recurred. Recurrence of macular oedema was defined as a decrease in visual acuity associated with an increase of intraretinal or subretinal fluid as detected upon optical coherence tomography or fluorescein angiography. The interval between the first injection and repeated injections was at least 3 months for triamcinolone and it was at least 6 weeks for bevacizumab. There were 1.3±0.4 re-injection in the triamcinolone group (range: 1–2 injections), and 2.7±1.8 re-injection in the bevacizumab group (range: 1–6 injections).

Statistical analysis was carried out using a commercially available statistical software package (SPSS for Windows, version 16.0, SPSS, Chicago, IL, USA). Best-corrected visual acuity was converted into the logarithm of the minimum angle of resolution for statistical calculation.8 The data that were distributed normally were presented as the mean±SD. Where appropriate, the Student's t-test and the χ2-test were used. Confidence intervals were presented. All P-values were two-sided and were considered statistically significant when the values were less than 0.05.

Results

Both study groups did not vary significantly (P>0.15) in the preoperative data (Table 1). The mean follow-up was 7.8±4.3 months (range: 3–12 months).

In both study groups, the mean visual acuity increased significantly (P<0.001) from baseline during the follow-up. In the triamcinolone group, the differences were significant for the comparisons between the baseline examination and the follow-up examination carried out at 2-months (P=0.03) and 3-months (P=0.02) after baseline (Figure 1). In the bevacizumab group, the differences were significant for the comparisons between the baseline examination and the follow-up examination carried out at 1 month (P=0.03), 6 months (P=0.04), and at 1 year (P=0.04) (Figure 1).

Figure 1
figure 1

Mean best-corrected visual acuity in LogMAR and CMT from baseline to 1-year follow-up. CMT: central macular thickness; **: the P-value of comparison between CMT of two groups was lower than 0.01.

Full size image

Comparing both study groups with each other, the differences in the gain of mean visual acuity were not statistically significant (P>0.40) at any time during follow-up examination (Table 2). In a similar manner, the percentage of patients who improved in best-corrected visual acuity by ⩾2 lines or who lost in best-corrected visual acuity ⩾2 lines was not significantly (P>0.30) different between both groups (Table 2). Correspondingly, the percentage of patients who improved in best-corrected visual acuity at 3 months follow-up by ⩾3 lines (19 (45%) patients in the triamcinolone group, 10 (33%) patients in the bevacizumab group; P=0.31) or who lost in best-corrected visual acuity at 3 months follow-up ⩾3 lines (2 (5%) patients in the triamcinolone group, 1 (3%) patients in the bevacizumab group; P=1.0) was not significantly different between both groups.

Table 2 Change in best-corrected visual acuity in patients with non-ischaemic central retinal vein occlusions treated by intravitreal injections of triamcinolone (4 mg) or bevacizumab (1.25 mg)
Full size table

The mean macular thickness decreased significantly (P<0.001) in the triamcinolone study group from baseline to any re-examination carried out between 4 weeks and 1 year after the initial injection (Table 3). In the bevacizumab group, the reduction in central macular thickness was statistically significant for the follow-up examination carried out at 1 month, 2 months, and at 1-year follow-up (Table 3). The reduction in the macular thickness was significantly (P=0.006) more pronounced in the triamcinolone group than in the bevacizumab group at the 6 months follow-up examination (Figure 1).

Table 3 Central macular thickness at baseline and during follow-up in patients with non-ischaemic central retinal vein occlusions and treated by an intravitreal injection of triamcinolone (4 mg) or bevacizumab (1.25 mg)
Full size table

In the triamcinolone group, two eyes (5%) developed iris neovascularization at 6 months after the initial injection, and two eyes (5%) developed a vitreous haemorrhage at 3 weeks and at 7 months after the initial injection, respectively. In the bevacizumab group, one eye (3%) developed iris neovascularization at 3 months after the initial injection and one eye (3%) developed a vitreous haemorrhage at 6 weeks after the initial injection

In the triamcinolone group, intraocular pressure readings higher than 21, 30, 35, and 40 mmHg, respectively, were measured in 18 eyes (43%), 6 eyes (14%), 4 eyes (10%), and 1 eye (2%), respectively, whereas in bevacizumab group, the intraocular pressure did not vary significantly between the examination at baseline and the examination during follow-up.

Discussion

Within the last 6 years, intravitreal triamcinolone has widely been used for the treatment of intraocular proliferative, oedematous, and neovascular diseases including CRVO.6, 7, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 A disturbed balance of angiogenic and inflammatory cytokines has been reported to be associated with retinal vein occlusion,20 and experimental investigations and clinical studies have suggested a temporary antioedematous and antiangiogenic effect of intravitreal triamcinolone in eyes with CRVO.6, 7, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 The two major side effects of the intravitreal triamcinolone were a steroid induced increase in intraocular pressure and development of cataract.21, 22, 23, 24, 25 In contrast, studies on intravitreal bevacizumab by Rosenfeld et al and other researchers showed an improvement in visual acuity, reduction in macular thickness, and only minor complications in patients with CRVOs,26, 27, 28, 29, 30, 31, 32, 33, 34 so that intravitreal triamcinolone was rapidly exchanged by intravitreal bevacizumab for the treatment of CRVO. It agrees with this study, in which the best-corrected visual acuity improved significantly in the triamcinolone group and in the bevacizumab group with a no statistically significant difference in the gain in visual acuity between both study groups, although the reduction in macular oedema was slightly more pronounced in the triamcinolone group (Table 2) (Figure 1). The side effects in terms of an elevation in intraocular pressure were present in the triamcinolone group only, in a similar frequency and amount as already reported for Caucasians and in another study on Chinese patients.21, 22, 23, 24, 25

The finding of a discrepancy between a recurrence of macular oedema and continuously improved visual acuity agrees with an observation by Kriechbaum et al,32 in which 3 months after the injection macular oedema recurred and visual acuity remained unchanged.

The studies on the intravitreal use of bevacizumab for treatment of CRVO partially differ in the frequency of the bevacizumab application. In study by Kriechbaum et al,32 three initial injections were administered at 4-week intervals,32 although the intervals in other studies were usually 6 weeks to 2 months.34 In the investigation by Hsu et al,29 Iturralde et al,27 and in our study, only one initial injection was primarily given.

This study also agrees with a very recent investigation by Wu et al35 who compared intravitreal triamcinolone with intravitreal bevacizumab for treatment of macular oedema because of CRVO. The researchers concluded that intravitreal injection of triamcinolone or bevacizumab can both lead to a significant improvement in visual acuity and a resolution of macular oedema in patients with CRVO. However, the significant effect was not permanent. The efficacy of intravitreal triamcinolone acetonide showed no significant differences compared with intravitreal bevacizumab but seemed to cause more adverse events than bevacizumab.

There are limitations of our study. It is a hospital-based study so that without doubt a bias by the referral of patients was introduced. The self-reported duration of the symptoms was relatively long so that the results of our study may not be transferred for a fresh CRVO. Another weakness of our study was that some patients had undergone a retinal laser coagulation prior to be included into the study. The percentage of patients with a previous retinal laser coagulation was, however, not significantly different between both study groups (Table 1). Nonrandomization of the patients between the two study groups is another important limitation of our study. It was, however, the decision of the attending retinologist whether triamcinolone or bevacizumab was injected, with the same retinologist using the same drug during the whole study period. As the patients were randomly referred to the retinologists participating in the study, this flaw in the study design may not have markedly influenced the results of the investigation. Accordingly, the two study groups did not differ statistically significantly in their baseline parameters. Another limitation of the study is that intravitreal triamcinolone may have increased the formation or progression of cataract,22, 25 so that a vision-reducing effect of progressive of cataract might have hidden parts of a vision-improving effect of triamcinolone. Another limitation is the relatively long duration of symptoms before the treatment was carried out. The results of our study may, therefore, not directly be transferred to patients with a fresh onset of a CRVO. The relatively long duration of the symptoms before therapy was started may also be a reason why the macular thickness did not return to normal levels in all patients treated. Finally, one may consider that the patients were re-treated first when macular oedema returned. It differs from other treatment strategies in which three initial injections of bevacizumab are given in an interval of about 6–8 weeks. The treatment strategy in our study may, therefore, have led to an undertreatment.

In conclusion, in long-standing non-ischaemic CRVO, intravitreal bevacizumab and intravitreal triamcinolone are both associated with a comparable gain in visual acuity, although the reduction in macular oedema was more marked in the triamcinolone group. In view of the potential complications of intravitreal triamcinolone with respect to intraocular pressure elevation and cataract formation, bevacizumab may be preferred compared with triamcinolone for intravitreal use in non-ischaemic CRVO. If, however, intravitreal bevacizumab did not lead to an improvement in visual acuity, intravitreal triamcinolone may be tried, as a recent study suggested that in eyes with non-ischaemic CRVO in which intravitreal bevacizumab failed to improve vision intravitreal triamcinolone may lead to an increase in visual acuity.36

Summary