Sir,

We have read with interest the comments provided by Smith relating to our study. We appreciate his remarks and his interest in our article.

As required by the author, we will explain the reason why our results differ from previous literature on the rate of postoperative vitreous cavity haemorrhage (POVCH).

In the quoted paper by Yang et al, preoperative bevacizumab (1.25 mg/0.05 ml) was combined with C3F8 10%.1 This gas is a long-acting gas with haemostatic effect, which could explain the absence of early POVCH. In our series, we did not use any gas at the end of the surgery.

Our results also differ from those reported by Yeoh et al by way of substantial differences in the study settings.2 First, our inclusion criteria were presence of vitreous haemorrhage (VH) with active proliferative diabetic retinopathy, whereas their inclusion criteria were tractional retinal detachment (TRD) involving the macula with active neovascularization, rubeosis with VH and high-risk features for developing rubeosis. Second, all 31 eyes in our study were filled with air, in contrast with only 1 of the 10 eyes in their study (the other nine being filled with SF6 or water).

In our study, the amount of VH was graded using slit-lamp biomicroscopy from grade 0 to grade 3. This allows a better understanding of the severity of POVCH. Looking carefully at our results, it may be inferred that we reported the presence of a low grade of VH (grade 1: mild VH, fundal details possible/hazy view) only within the first week after the surgery, whereas grade 3 VH (severe recurrent VH with no fundus details) was present only in 3, 3, 6, and 6% of the eyes at 7 days, 1-, 3-, and 6-month follow-up, respectively.

The author has asked for clarification on the surgical technique.

The fluid–air exchange was performed at the end of the surgical procedure, whereas gas tamponade or silicon oil was never used. In our series, 28 out of 31 patients (90%) had already undergone panretinal photocoagulation earlier. The endolaser photocoagulation was carefully completed in all eyes during the surgery.

We did not find any toxic effect of IVB (2.5 mg/0.1 ml). The only adverse event was one single case of TRD that developed during the follow-up period.

In our opinion, the IVB injected at the end of the surgery does not have any effect on late re-bleeding, the main reasons for recurrent VH being intraoperative surgical manoeuvres, insufficient haemostasis, and persistent retinal neovascularizations.

Our hypothesis is that preoperative bevacizumab injection facilitates surgery by enabling a complete endolaser photocoagulation treatment, reducing intraoperative haemorrhage, and avoiding the need for a complicated manoeuvre to achieve total delamination of the fibrovascular tissue, thereby minimizing the risk of recurrent vitreous haemorrhage.

Conflict of interest

The authors declare no conflict of interest.