The technique of the patch operation has been developed further by Jan van Meurs, who has documented the functionality of the free transplant in large case series. His technique is low risk in that it is little different from submacular membrane extraction. He refrains from 180° peripheral retinotomy and from the detachment of the temporal retina, and implants the transplant by the same paramacular retinotomy through which he previously extracted the submacular CNV [11].

In spite of the obvious advantages of the patch method compared to macular translocation, this method has not yet been generally accepted, even by experienced retinal surgeons. The basic reason is that most surgeons choose access to the subretinal space via a peripheral 180° retinotomy and get bad functional results. Insertion of the patch through a paramacular retinotomy, as practiced by van Meurs, is complex. Access to the submacular space is limited. For instance, it is not possible to stop bleeding underneath the macula. Bleedings originating from the choroid happen from time to time, be it during CNV extraction, through imminent massive subretinal bleeding or postoperative secondary bleeding. An extensive diathermia underneath the macula would be desirable in order to avoid secondary bleedings, especially as many patients of this age group take coagulation-inhibiting medication. Ultimately, it is unavoidable during massive submacular bleedings to combine the patch method with peripheral 180° retinotomy and folding over of the retina, as it is otherwise not possible to remove massive bleedings up to the retinal periphery from below the retina.

The author practices both methods and confirms that the functional results of the free transplant with 180° retinotomy, compared with the successes that we are used to from paramacular access, have not met our expectations. The reason why this is the case is currently under discussion. One difference between both methods is the application of liquid perfluorcarbon (PFCL). The folded-over retina becomes immobilized by PFCL. PFCL comes into contact with the photoreceptors. Indeed, the heavy liquid gets drawn off, but we know that a thin PFCL coating on the retina – in this case the photoreceptor layer – remains. In another context, we know from subretinally remaining PFCL that it damages the retina. The possible retinotoxic effect of PFCL on photoreceptors is currently being investigated (Robert McLaren, London, personal communication). So far, we could not do without

PFCL as shifting the transplant from peripheral to central is not controllable without the heavy liquid.

18.5Indications for Surgery

Looking at the results of the ranibizumab approval studies over 2 years and the low risk of intravitreal injection, it makes sense to reduce the surgical approach to those indications that are not covered by VEGF-blockers. Low-risk treatment allows for an early start. However, the few (infection) risks become practically significant should injections be necessary in short intervals over many years. However, should it become apparent during the approval studies that the mean visual acuity cannot be maintained in practical application within 1 year [12, 13] the surgical indication may expand again in the future. The patch method records that the visual level remains stable as of the 6th postoperative month of an observation period of 1 year [14].

18.5.1 Non-responder

The “non-responder” is not uniformly defined for VEGF-blockers. Approval studies suggest a quotient of 5%, which shows a deterioration of vision after the first loading phase. As it is sometimes possible to reach an inactive state after the second or more anti-VEGF treatment series, further monthly injections are performed. Other alternatives (photodynamic therapy) are tested. Only when vision is very bad, effort and costs of the intravitreal injection are compared with the last limited visual prognosis, and then lasting alternatives of treatment are considered. In these advanced stages, the aim of surgery can only be stabilization on a low level.

18.5.2 Pigment Epithelium Rupture

The pigment epithelium rupture is a complication of the pigment epithelium detachment. It is a relatively urgent incident. This means a chance of vision improvement when treatment starts early. When the fovea is no longer supported by the pigment epithelium and shows no retinal edema, stabilization cannot be expected by means of a VEGF-blocker or spontaneously. It has been unsuccessfully tried to unfold the curled-up RPE in

Fig. 18.1 Pigment epithelium rupture before and 3 and 10 months after patch. (a) Preoperatively, microperimetry shows no response in the upper half of the macula and in the temporal lower macula quadrant. Postoperatively, macula function recuperates in the upper half. (b) OCT preand postoperatively.

Preoperatively, the macula shows no edema. The RPE is retracted so far that it cannot support the fovea. Postoperatively, the macula lies smoothly on the transplant. The foveal depression is maintained. Reading acuity of 0.4 preoperatively recovers to 0.6 and stays stable

such a way that the fovea continues to be supported. Thus, the autologous transplant (or macular translocation) is a sensible therapeutic option, which even leads to an improvement of vision when started early (Fig. 18.1).

18.5.3 Massive Submacular Bleeding

Massive subretinal bleeding is a complication of exudative AMD. It happens disproportionately often when AMD patients take coagulation-inhibiting medication [15]. Submacular bleeding manifests itself in manifold ways.

•Bleedings can be restricted to the macula but can also extend beyond the temporal vessel arcades, in an extreme case up to the peripheral retina. The latter results in the loss of ambulatory vision.

•The macula is omitted by the bleeding. This is an indication that the CNV is pathologically adherent to the retina and indicates a poor prognosis.

•The bleeding can remain restricted to the compartment underneath the pigment epithelium (hemorrhagic lifting of pigment epithelium). Sub-pigment epithelial bleedings become clinically apparent when they spread subretinally as well. The patient either

sees a central scotoma or – with previous central scotoma – a sudden enlargement of the scotoma. According to the principle “as little as possible, as

much as necessary,” the surgical indication is gradually adapted according to the extent of the bleeding. When deciding on major and complex surgery, the higher risk of secondary bleeding should be considered. In detail there are the following options:

1.Intravitreal injection of rTPA and gas with positioning of the patient

2.Vitrectomy with subretinal rTPA injection and gas tamponade

3.Vitrectomy and subretinal extraction of smaller bleedings via paramacular retinotomy, if applicable, together with submacular membrane extraction and patch

4.Vitrectomy and subretinal extraction of larger bleedings through peripheral 180° retinotomy, if applicable together with submacular membrane extraction and patch or enlargement of the retinotomy to 360° with macula translocation

Subretinal bleedings are generally recognized imme-

diately, so there is a chance to start treatment straight away.

1. For intravitreal injection with rTPA and gas, a time span of up to 2 weeks after the start of bleeding is

generally considered effective. Depending on the extent of the bleeding, one has to weigh up whether this procedure is promising. The gas volume is relatively small, as is the displacement power produced by this gas bubble. Most suited are shallow bleedings. For lack of classified field reports, the surgeon decides as he sees fit.

2.A disproportionately larger and thus longer lasting gas volume can be put into the vitreous cavity in combination with vitrectomy. Compared to the above mentioned method (see 1.) [1], this expansion may allow repression of prominent bleedings beyond temporal vessel arcades. In spite of numerous reports, no consistent recommendations exist as to whether adjuvant administration of rTPA is at all necessary or whether it should be applied intravitreally or subretinally. Anyway, most surgeons use additional rTPA. Wilson Heriot [16] is convinced that the rTPA diffuses well through the retina so that no subretinal injection is necessary. Possibly, when injecting rTPA subretinally, additional liquid allows for an easier repression of the blood to the periphery. These first two options leave the CNV where it is. Administration of a VEGF-blocker is therefore indicated either combined or shortly after injection.

3.In contrast to the two options mentioned above, a third possibilty is to open the subretinal space and to remove the blood. The patient does not have to be positioned. In the end, the macula rests either on the previous pigment epithelium or on a free transplant of RPE and choroid. No published case series exists on this. Own experience shows that there is a high risk of fibrosis of the transplant through secondary bleeding when combined with a patch.

4.The peripheral 180° retinotomy allows optimal access to the subretinal space, so that, even if the CNV is older and therefore coagulated and fibrotic, bleedings up to underneath the peripheral retina can be removed. Since the retina can be folded over in the nasal direction, the choroid is easily accessible in the macula region. This eases the membrane extraction and facilitates an extensive coagulation as prophylaxis of secondary bleedings. When this type of surgical intervention is finished by using a free transplant, functional results are not satisfactory. The reasons for this are unknown. Indeed, the first good results with this operating technique were reported by Barbara Parolini (2010) [10]. Many of the surgeons who are familiar with macula

translocation prefer to finish surgery with a macula translocation, even if it means an extension of the retinotomy to 360°. Initially, the bleeding conceals the extent of the pigment epithelium lesion. Whether surgery is finished with a patch, with macula translocation or without pigment epithelium replacement can only be decided during surgery.

18.5.4 Dry AMD

Atrophy of pigment epithelium without the complication of neovascularization can so far not be treated with medication. This should open the domain of surgical therapy, like in fact that of the free transplant (see Sect. 18.2). Forty percent of blindness and serious visual impairment within AMD is caused by geographic atrophy. It is technically possible to transplant pigment epithelium and choroid into the previously damaged submacular space [17]. One may expect this type of surgery to be less complex than surgery for exudative AMD as there is no membrane extraction. However, surgery of dry AMD places special demands on the surgeon:

•If one looks – for example, in high resolution OCT - at the pathologically tight adherence of the macula to Bruch’s membrane and choroid, it is not surprising that the lack of exudation makes it rather difficult to lift the macula without additionally and irreversibly traumatizing the already compromised outer retina. When injecting BSS into the subretinal space, at first the retina often detaches outside the macula. There may possibly be no alternative but to displace the macula from its base with a potentially traumatizing spatula maneuver.

•One cannot dismiss the lesion of Bruch’s membrane, including the bleeding risk that automatically occurs during membrane extraction. An intact Bruch’s membrane means an effective barrier for vessel anastomoses, also for the reperfusion of the transplant (Fig. 18.2). Additionally, these experiences signal that free transplants without choroid do not function. Contrarily, the importance of the integrity of Bruch’s membrane becomes evident in that a CNV can be induced also in dry AMD as soon as Bruch’s membrane is damaged (Fig. 18.3).

Due to the slow progressive atrophy of the macula

prior to transplantation, a visual improvement is usually impossible. Through transplantation, an eccentric

Fig. 18.2 (a, b) Pigment epithelium choroid translocation with dry AMD. Here, the aim was not to open Bruch’s membrane. Postoperatively, the patient notices a dark central scotoma. (a) 6 months after the operation, the ICG angiography shows no vessels. The transplant appears as a black rectangle. Underneath,

Fig. 18.3 Fluorescein and ICG angiography (a, b) of a CNV at the left lower edge of a pigment epithelium choroid transplant. The complication arose postoperatively with dry AMD. Regeneration of the vessels signals that a geographic atrophy can be accompanied by a clinically relevant VEGF level in the eye. Only the

the local choroid vessels are visible in a zone of pigment epithelium atrophy. (b) In spite of absent revascularization, the autofluorescence of the transplant is maintained 6 months postoperatively. (c) ICG angiography of a perfused transplant of a different case in comparison to (a)

intraoperative lesion of the Bruch’s membrane enables the transition into an exudative course of progress. Since the CNV lies on the edge of the transplant and is thus extrafoveal, it can be conventionally thermolaser-coagulated

fixation cannot be changed back to a central fixation. If surgery may only achieve a stabilization of the situation, it only makes sense if there is suffcient vision left worthwhile preserving. Considering the complication rate, the best time of surgery is the impending loss of reading vision without the use of a magnifying glass. At this stage, the psychological strain of the patient is low, especially when only one eye is affected. Even so, due to the above-mentioned surgical risks,

the risk of iatrogenic deterioration of vision is high (30%) [18].

All these considerations complicate recommendations for surgery in individual cases. Our own results and experiences of the London group [19] concerning transplantations in patients with geographic atrophy and macular dystrophy teach us that even with successful surgery (stable vision) patients are less satisfied than with surgery for exudative AMD. This is the