**4.4 Special considerations in CRD**

Most often, one may not be able to identify a break pre-operatively. Convex configuration of detachment, SRG, subretinal haemorrhage are indirect clues towards CRD.


Apart from truncating the cone, identifying the right plane and second membranes, dissection of fibrovascular proliferation, another major challenge in diabetic vitrectomy is controlling the haemorrhage or clot management. Bleeding can occur from multiple sources, i.e., vascular nails during FVP dissection, edges of FVP during dissection, optic nerve head due to dense adhesion of posterior hyaloid or FVP at the disc, arcade vessels due to iatrogenic damage, edges of retinal breaks or trimmed NVEs. Despite multiple techniques to arrest the bleeding, it may sometimes be very difficult to handle these blood clots as they get densely adherent to the underlying thin retina.

### **4.5 Management of haemorrhage**

This is one of the crucial and sometimes most challenging step during diabetic vitrectomy. Clot management should always be preferred in the fluid filled cavity, as the air would cause the clot to diffusely get adhered to the retina and will be more catastrophic. Once all the bleeders are managed, before switching to air, one can decrease the IOP to look for any residual oozing and manage accordingly. This would help to reduce post-operative bleeding.

In case of an active visible bleeder as seen from either the edge of FVP or NVEs, one can immediately cauterise them using endodiathermy. In cases where the bleeding is from a major arcade vessel, one has to be cautious to use diathermy as it may cause vascular occlusion subsequently. Light burns using endolaser probe (Power: 150–200 mW and shorter duration—100 ms) can be attempted instead. Rarely, one can also try pinching of the vessel gently using forceps as a last resort.

In case of bleeding from the disc, we cannot use diathermy or endolaser over the disc. Hence either increasing the bottle height of infusion (in case of gravity assisted infusion) or temporarily increasing the intraocular pressure may help to arrest the bleed. If nothing works, a fluid air exchange can also be tried to use air as a temporary tamponade and wait for few minutes to arrest the active bleeding. One has to be patient and may need to make repeated attempts of the above manoeuvres to achieve haemostasis. One does not have to always peel the membrane over disc, as it may bleed unstoppably if it is densely adherent and vascular. This can be circumcised or trimmed and left behind.

Not always, every blood clot needs immediate attention. Sometimes if it has clotted and is not hampering further membrane dissection, then the clot may be left alone and addressed at the end once all membranes are dissected. Small clots can be removed using flute, while larger clots may need cutter. Clots not covering the macula can be left alone if there is no surrounding break or traction if they cannot be safely removed.

#### **4.6 Vitrectomy in tractional versus non-tractional DME**

Patients with tractional macular edema in diabetes are known to benefit with vitrectomy and ERM with ILM peeling [25]. In non-tractional DME, role of

#### *Vitrectomy in Diabetic Retinopathy DOI: http://dx.doi.org/10.5772/intechopen.101358*

vitrectomy has been controversial. While some studies show that patients with refractory DME would benefit with vitrectomy [26], other studies have shown no significant visual improvement with ILM peeling although there may be anatomical improvement [27, 28]. Addition of intravitreal steroid along with ILM peeling in some cases of refractory DME, has shown to improve visual outcome in longterm [29].

In patients with long standing macular ischemia or FVP causing traction over fovea, sometimes a macular hole may be noted. Visual prognosis in such cases is guarded despite ILM peeling in contrast to idiopathic macular holes [30].
