Article Text

Download PDFPDF

Early chorioretinal anastomosis in non-ischaemic CRVO: a randomised trial
  1. R J Antcliff,
  2. E J Mayer,
  3. T H Williamson,
  4. J S Shilling
  1. Department of Ophthalmology, St Thomas’s Hospital, Lambeth Palace Road, London E1 7EH, UK
  1. Correspondence to: Eric Mayer University of Bristol, Clinical Sciences, Bristol Eye Hospital, Lower Maudlin Street, Bristol BS1 2LX, UK;

Statistics from

In non-ischaemic central retinal vein occlusion (niCRVO), the two principal determinants of final visual acuity are visual acuity and the presence of macular oedema at initial presentation.1,2 Data from the Central Vein Occlusion Study Group2 suggested that of patients with CRVO with an initial acuity better than 6/15, 65% maintain this, whereas patients with presenting acuities between 6/15 and 6/60, 81% remain the same or get worse (19% improve to 6/12 or better, 44% remain between 6/15, and 6/60 and 37% are worse than 6/60). Up to 34% of niCRVO may progress to the ischaemic variant with its attendant complications within 3 years.2 However, in some studies 83% of indeterminate cases progress to ischaemic CRVO.1 The creation of a chorioretinal venous anastomosis (CRVA) improved visual acuity in some patients, decreased macular oedema, and reduced the incidence of progression to ischaemic CRVO.3–5 These important observations provide the basis for the study we carried out.

During the study period 11 patients (table 1) were enrolled according to the trial protocol (see appendix). All anastomoses were patent (on fluorescein angiography and if not repeated, see appendix). Mean patient age and mean pretreatment (T0) visual acuities, retinal thickness (by optical coherence tomography, OCT), and cyst height (OCT) did not differ between the two groups (table 1, fig 1). Factorial analysis of variance (ANOVA, Genstat) found that all three measures (acuity, retinal thickness, and cyst height, fig 1) decreases significantly between T0 and 6 months, but only in the case of cyst height was there a significant difference between the laser treated and control groups (F(1,9) = 5.85, p<0.05). The changes in retinal thickness and visual acuity were also greater in the treatment group but did not achieve significance (F(1,9) = 2.51 and 1.15, respectively, both p>0.1). Larger group sizes may have shown a significant effect.

Table 1

 Comparison of randomised groups for early laser chorioretinal venous anastomosis versus observation in non-ischaemic CRVO (patient data are illustrated, visual acuities are given as ETDRS measures)

Figure 1

 Data of retinal thickness (A), cyst height (B), and visual acuity (C) for patients randomised to laser chorioretinal venous anastomosis or observation. Visual acuity is measured on the ETDRS scale. Retinal thickness and cyst height were measured in pixels (1 pixel = 4 μm). Error bars show standard errors of means, obtained from factorial analysis of variance. Significant effects are identified as follows: *p<0.05; **p<0.01.

The principal limitation of this trial is the small number of patients, because of the limited numbers of suitable patients within the study period. However, patients were randomised and well matched. The trial showed a trend towards better vision over 6 months in the laser treated group, although this failed to reach statistical significance. It remains possible that a larger group of randomised patients would have produced statistically significant numbers. A multicentre trial would be required to achieve this. Another question arises regarding the longer term outcome in these patients. Again this should probably be considered in further studies. Cyst height might predict later visual acuity as there is evidence that it correlates with visual acuity outcome measures.6,7 It is therefore possible that longer follow up may have found a significant improvement in acuity. Retinal thickness predicts and precedes changes in visual acuity in diabetic retinopathy.8

If the size and duration of this trial had been increased, it is conceivable that results may have proved more informative. The fact that there was a significant effect on cyst height on OCT is, nevertheless, important and suggests that laser CRVA may reduce macular oedema in CRVO. The finding that improvements in visual acuity are recorded later than reduction in macular oedema8 in other studies implies that a significant effect on visual acuity might have been expected in these patients at a later time. We therefore recommend that this be taken into account in future studies.


The principal study end point was visual acuity at 6 months and resolution of macular oedema, the secondary end point was the identification of subgroups on optical coherence tomography (OCT 2000, Humphrey Instruments, USA) with a more favourable outcome including reduction of consecutive ischaemia. For entry into the study patients were required to: (1) be within 1 month of symptom onset; (2) have visual acuity of 6/24 or worse; (3) have good retinal perfusion (less than 10 disc diameters of non-perfusion); (4) have no relative afferent pupillary defect or neovascularisation; (5) have macular oedema confirmed by optical coherence tomography and fluorescein fundus angiography (FFA). The following exclusion criteria were applied: diabetes with macular oedema in the fellow eye; inability to give informed consent; pregnancy; sensitivity to fluorescein; cloudy media sufficient to preclude adequate fundal photography; age less than 40 years.

Upon entry into the trial the following were performed: best corrected visual acuity (Snellen and ETDRS); fundus photography, and fluorescein angiography; OCT assessment of the macula. Patients were then randomised (envelope) to laser or observation groups. Laser anastomosis was performed with a 50 μm spot size for 0.1 second with 1.5–2.5 W (maximum power possible) with an argon laser. The first shot was fired at the edge of the vein to rupture Bruch’s membrane and the second shot at the edge of the vein to rupture the vein. A third shot with a YAG laser was applied (3–5 mJ) if required. Anastomotic sites were attempted in the following locations in order of preference—inferonasal, inferotemporal, or superonasal. All anastomoses were made at least 3 disc diameters from the optic disc. At 3–4 weeks repeat FFA was performed to determine anastomotic success: where no anastomosis was present this was repeated, until functioning.

Patients were reassessed at 1, 2, 3, 6, and 12 months and acuities and OCT recorded.

It was expected that approximately 40 patients would be required to show a 50% benefit.

Because of laser surgery it was not possible to mask the patient or observer as to which treatment had been carried out.

No adverse events were reported or noted in this series.


We gratefully acknowledge the help of Professor Stephen Dunnett (Cardiff University) with statistical analysis of the data.



  • RJA was supported by the Lady Allerton Fund.

  • Competing or financial interests: none.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.