We now focus on vitrectomy maneuvers. The purpose of this presentation is more to describe which techniques are currently used among the surgeons community and to see what are the consequences on the failure rate.

For PVR stage B and C1, most surgeons have recourse to posterior hyaloid removal and it’s seems reasonable since without posterior hyaloid removal, effective failure rate is quite high : 7.1%. The difference of this failure rate with posterior hyaloid removal is statistically significant.

As far as ILM removal is concerned, for PVR stage B and C1, without choroidal detachment and hypotony, there are about three times more surgeons who do not perform any ILM peeling. No statistically significantly difference between with and without ILM peeling with respect to failure rate.

All PVR stages taken together, but choroidal detachment and hypotony excluded, it appears that not performing any retinotomy leads to better results as much concerning the true failure rate as concerning the remaining silicone rate.

However, when studying the results for each PVR stage, this trend remains significant only for stage C1. One more time PVR stage C1 appears to be very different from the other PVR stages and therefore it will be treated separately from the other PVR stages concerning the maneuver.

The previous multivariate analysis had showed the influence of retinotomy.

The retinopexy distribution shows that the great majority of surgeons use either either to 360 laser, either focal laser or cryo possibly reinforced or not by Focal laser.

When studying this distribution according to PVR stage, it’s noticed that the trend is the following : when the PVR gets worse, it is used, in proportion, less cryo or focal laser and more 360 laser.

At first sight, cryo is more efficient than focal laser and 360 laser since its failure rate is statistically lower, however,

cases treated with 360 laser have more advanced stages of PVR than cases treated with cryo or focal laser, which can explain the difference in outcomes previously seen.

So that only the use of Cryo versus Focal laser is a major independent explanatory variable of failure rate.

When a vitrectomy is performed, there are only a very few cases for which no tamponade is used.

The distribution of the tamponade used is quite interesting since it shows well that in more advanced stages of PVR, the more silicone is used. This trend is once again particularly obvious for PVR stage C1.

When looking at the failure rates for gas and silicone tamponade, it is logical that silicone tamponade leads to an incredibly higher remaining silicone rate than gas tamponade. The rate of remaining silicone with a gas tamponade is however not 0, which means that the eye was re operated after the primary procedure. As far as the effective failure rate is concerned, silicone tamponade appears to be less effective than gas tamponade, but keep in mind that this results concern all the PVR stage grouped together.

The PVR distribution according to the tamponade type demonstrates that there is a much more advanced degree of PVR in cases treated with oil.

Therefore in order to get a better comparison between gas and silicone tamponade, it is necessary to compare the failure rates at each PVR stage. The choroidal detachment and hypotony were also excluded. It then appears that the supremacy of gas over silicone is particularly obvious for PVR stage C1.

So that the tamponade type is a major independent explanatory variable of failure rate.

One last remark that can be made about tamponade is the distribution according to the pump used. It is particularly obvious that users of a peristaltic pump do not show same behavior regarding tamponade type. Venturi users use silicone oil more frequently than peristaltic users.

Finally, we can see that a majority of surgeons do not perform any buckling when they perform a vitrectomy.

It seems that in proportion more buckles are performed at advanced PVR stages, however the trend is not really obvious.

According to these failure rates, buckling with vitrectomy certainly does not seem to offer any advantages - on the contrary.

Once again, we can see that there is a visible difference in strategy between peristaltic users and venturi users. A segmental buckle is very rarely performed by users of venturi pump and quite often by those with a peristaltic pump. There is a real difference in strategy based on the pump owned.

A multiple correspondence analysis was performed in order to see if there were groups with common characteristics based on vitrectomy and external techniques. What we had guessed thanks to the graphics you have just seen was confirmed : when a vitrectomy is performed, it is often one of the two following combinations that is performed : Perist + low cutting speed+gas+segmental tamponade or Venturi pump + high cutting speed+silicone tamponade+360 buckle.