Eye (1989) 3, 277-281

Indentation Microsurgery: Internal Searching for Retinal Breaks

P. H. ROSEN, H. C. WONG and D. McLEOD London

Summary Breaks responsible for rhegmatogenous retinal detachments in 78 eyes could not be seen preoperatively owing to opacities in the media, previous buckling or other causes. Deep kinetic indentation of the sclera combined with endoillumination enabled retinal breaks to be identified during closed microsurgery in 95% of these eyes, and retinal reattachment was eventually achieved in 85%.

Gonin stated the principles of repair of rheg­ Wong, Billington and Chignell reported matogenous retinal detachment, viz identi­ encouraging results of both 'speculative' con­ fication of the retinal breaks and their ventional retinal surgery and an 'internal' effective closure.! However, difficulties may vitrectomy approach for eyes in which retinal arise in locating retinal breaks owing to breaks could not be seen before surgery. 2 opacities in the media (including problems of However, breaks were discovered in only 18 pupil dilatation), or the position and/or size of of the 47 vitrectomy patients. We describe an the breaks. These factors are inter-related; for indentation microsurgery technique which, in example, large retinal breaks can generally be our hands, has greatly increased the chances seen by ophthalmoscopy despite considerable of identifying retinal breaks (especially those obscuration in the media, while small breaks located near the ora serrata), and we report may prove impossible to locate even if the the outcome of such surgery in 78 eyes. media are perfectly clear. 2 Lincoffand Geiser3 have shown that the distribution of subretinal Material and Methods fluid (SRF) can be a useful predictor of the We reviewed a consecutive series of patients position of retinal breaks, but "speculative" who underwent pars plana vitrectomy for buckling procedures based on these guide­ rhegmatogenous retinal detachment between lines in eyes with unidentified breaks are of January 1985 and October 1987 in whom the only limited value,2A especially in cases of principal reason for undertaking vitrectomy failed previous retinal surgery, total or subto­ was to permit identificationof the responsible tal detachment or major obscuration of fun­ retinal breaks. We excluded from consider­ dus details. Alternatively, a closed ation eyes with combined traction and rheg­ microsurgical technique can be employed in matogenous retinal detachment associated order to remove any media opacities and thus with trauma or vasoproliferative retinopa­ to permit identification of the breaks.2•s By thies (including diabetic retinopathy), this means, elements of vitreoretinal traction together with eyes with proliferative can also be managed and space created for vitreoretinopathy (PVR) in which the main intravitreal injection of gases (or silicone oil) reason for recourse to vitrectomy was removal for postoperative internal break tamponade. of tractional membranes rather than identi-

From Surgical Vitreoretinal Unit, Moorfields Eye Hospital, London.

Correspondence to Professor D. McLeod, FRCS, Royal Eye Hospital, Oxford Road, Manchester M13 9WH. 278 P. H. ROSEN ET AL. fication of breaks. Seventy-eight eyes of 78 mobile retinal detachment was made by patients were included in this series; the B-scan ultrasonography. patients ages ranged from nine to 88 years The following were the reasons for resort­ (mean 51 years). Of these, 40 eyes were pha­ ing to an internal search, either singly or in kic in the visual axis, 27 were aphakic, nine combination, in the 78 eyes in this series: were pseudophakic and two had dislocated opacity of the media (56 eyes), including lenses. Axial myopia of more than -6.0 vitreous haemorrhage (12 eyes), vitreous pig­ dioptres (or its aphakic equivalent) was pres­ ment and debris (30 eyes), lens remnants ent in 28 (36%) of eyes. Scleral buckling or (12 eyes), cataract (10 eyes) or corneal encircling procedures had previously been opacities (3 eyes), pupillary miosis or distor­ undertaken (unsuccessfully) in 48 eyes (62%) tion (20 eyes), deep previous scleral indenta­ on up to five occasions (mean 1.6 failed pro­ tion (12 eyes), and highly bullous retinal cedures in this 'secondary' group). The retina and/or choroidal detachment (7 eyes). was totally detached in 35 eyes, subtotally detached in 24 eyes (i.e. less than one quad­ Preliminaries to Internal Searching rant attached), or less extensively detached in Patients were carefully positioned for surgery 19 eyes; in 63 eyes the macula was detached. so that all quadrants of the fundus were Tangential traction from epiretinal membrane accessible for internal searching: in essence, contraction (PVR grade B or C1) was evident the head was secured so that the coronal plane in 13178 eyes (17 per cent). of the eye was horizontal but with a slight tilt nasally to improve access to the nasal periph­ ery. Subconjunctival mydricaine was given Problems of Fundus Visualisation routinely to ensure maximal pupillary dilata­ In 71 out of the 78 eyes (constituting category tion throughout the procedure. A), the retina was identifiedas being detached Vitrectomy was performed using 20-gauge by ophthalmoscopy but no retinal breaks were instrumentation and a standard 3-port system. seen preoperatively or, if a break was seen The infusion cannula was generally placed (17 eyes), it was not thought to be primarily or opposite the lower border of the lateral rectus solely the cause of the detachment by virtue of insertion to facilitate temporal rotation of the its site in relation to the configuration of the globe; a modified infusion cannula (Fig. 1) detachment. In the remaining seven out of was sometimes employed either inferotem­ 78 eyes (category B), neither retinal breaks porally or inferonasally, and on rare occasions nor the retinal detachment could be seen a lateral canthotomy was necessary. The ophthalmoscopically, and the diagnosis of active entry sites for endoillumination and vitrectomy were made opposite the upper borders of the horizontal recti in order to max­ imise the angular range of internal inspection and surgery. Scleral indentation facilitated removal of the basal vitreous gel following posterior vitrectomy, and vitrectomy was combined with other surgery to improve view­ ing of the posterior segment in 23 eyes (lens extraction in two eyes; lensectomy in nine eyes and anterior segment revision in 12 eyes).

Internal Searching Posteriorly The post-equatorial retina was inspected using endoillumination, and suction was applied by means of a flute needle or linear Fig. 1. Right-angled infusion cannula for internal aspirator to any areas suspected of harbouring searching. retinal breaks. Transretinal flow of SRF INDENTATION MICROSURGERY 279 enabled any full-thickness discontinuity to be retinal breaks at the edge of the optic disc; confirmed. three eyes with macular holes and adjacent epiretinal membranes; one eye with retinal Internal Searching Peripherally incarceration into a surgical entry site requir­ One of the two superior entry ports was plug­ ing extensive retinectomy; and one eye with ged and, with the height of the infusion bottle breaks in grossly atrophic retina and with a lowered, 1800 of peripheral retina was exam­ very thin sclera from multiple previous surgi­ ined by endoillumination combined with sur­ cal procedures. geon-controlled deep kinetic scleral Retinal breaks were treated by transcleral indentation (Fig. 2). By 'changing hands' i.e. cryotherapy in 49 eyes, by endolaser photo­ by plugging the other active port and replac­ coagulation in 23 eyes, and by both modalities ing the endoilluminator within the globe in fiveeyes. No retinopexy was carI'\ed out in through the other entry site, internal search­ one eye owing to peroperative choroidal ing was repeated throughout the remaining haemorrhage. ocular circumference. Except for one primary procedure for a localised retinal detachment owing to a single Retinal Reattachment Surgery small break, all eyes had a scleral buckle SRF was drained internally in 40 eyes (via a and/or encircling band over the vitreous base retinal break in 33 eyes or via a deliberate region at the completion of surgery. Equa­ retinotomy in seven eyes) and externally in torial breaks were supported by a break-ora seven eyes; both internal and external drain­ occlusive buckle.6 In the four eyes in which no age of SRF was performed in one eye. In the retinal breaks were discovered, an encircling remaining 30 eyes, fluidwas not drained from band was placed at the posterior border of the the subretinal compartment. vitreous base together with tyre explants over Air or an SF6: air mixture was exchanged particularly suspicious areas of the retinal for vitreal fluid in 70 eyes to allow post­ periphery. operative internal tamponade of breaks. In one eye, no tamponade was used owing to Postoperative Positioning peroperative choroidal haemorrhage. Sili­ For four or more days postoperatively, cone oil was injected into the remaining seven patients adopted a face-down or other appro­ eyes for the following reasons: two eyes with priate posture in order to maintain the arc of contact of the tamponade agent against the site of the retinal break. In eyes in which subretinal fluid was not drained at surgery, patients were alternately postured imme­ diately postoperatively in the right or left sem­ iprone positions at hourly intervals in order to prevent redundancy folding of the retina pos­ terior to the buckle.

Results I. Break Identification In 74 out of 78 eyes (95%), breaks were dis­ covered peroperatively which were felt to be responsible for the retinal detachment. This included additional breaks in all 17 eyes in which, although a break was seen pre­ operatively, it was not thought to be primarily responsible for the detachment.

Fig. 2. Drawing of internal searching using endo­ In 15 eyes the break was post-equatorial illu mination and surgeon-controlled deep scleral and in 52 eyes pre-equatorial; in seven eyes indentation. both pre- and post-equatorial breaks were 280 P. H. ROSEN ET AL. seen. In the remaining four eyes, no retinal (Table I) but three of the failures were cate­ break could be identified peroperatively. Of gory B patients. All eyes in which SRF was not the posterior breaks, five were macular holes drained peroperatively were successfully reat­ and three were breaks at the edge of the optic tached justifying this type of procedure in disc, while seven of the 22 eyes with post­ selected cases. equatorial breaks were highly myopic. The number of breaks identifiedin each eye was 1-8 (mean 1.9), and there was a substan­ IV. Retinal Reattachment at Six Months tial preponderance of peripheral breaks in the Postoperatively superior hemisphere. There was no difference Sixteen of the patients with a successful ana­ in the success of break discovery comparing tomical outcome at one month post­ categories A and B, or comparing primary operatively were lost to further follow-up and secondary procedures (Table I). because they returned home overseas. Of the In the 48 eyes with previous scleral buckling 62 eyes which were re-examined at six months or encirclement, the newly-discovered breaks (or which were known to be failures), 53 eyes were located anterior to the buckle in nine (85%) had complete retinal reattachment eyes, posterior to the buckle in 20 eyes, to the (Table I). Among these 53 eyes, however, ten side of the buckle in four eyes, on the buckle in had required further retinal reattachment sur­ two eyes and with no relation to a segmental gery in the interim (the retina remaining buckle in the remaining eyes. attached for a period of six months from the last surgical intervention). Thus 43 out of II. Peroperative Complications 62 eyes (70%) which were reviewed at six Significant peroperative complications arose months had an attached retina and had in seven of the 78 procedures: lens touch required no further vitreoretinal surgery after occurred in two eyes, sub retinal infusion in the internal search. Surgery was especially one eye, retinal incarceration in an entry site definitive in the 40 secondary procedures in in one eye and choroidal haemorrhage in which the retina remained attached through­ three eyes. out the six postoperative months in 32 eyes (80%). III. Retinal Reattachment at One Month Of the seven eyes with peroperative compli­ Postoperatively cations, five were successfully reattached at Sixty-seven out of 78 eyes (86%) had a com­ six months (the two failures having suffered pletely reattached retina at one month choroidal haemorrhage peroperatively). The following surgery. Within this period, how­ precise reason for failure was not always evi­ ever, three eyes had required further retinal dent in the nine eyes which had a persistently reattachment surgery. The anatomical success detached retina six months following internal of surgery was similar whether the internal searching. However, PVR was a prominent search was a primary or secondary procedure feature of most of these eyes, either as a cause

Table I. Break discovery and retinal reattachment after internal searching in 78 eyes

Retina reattached Retina reattached Break seen one month six months peroperatively postoperatively postoperatively

Overall 74/78 67/78 53/62 Category A 67171 63171 48/56 Category B 717 4/7 5/6 Primary procedure 28/30 25/30 17122 Secondary procedure 46/48 42/48 36/40 SRF drained 37/48 30/38 SRF not drained 30/30 23/24 Break seen peroperativcly 64/74 52/59 Break not seen peroperatively 3/4 1/3 INDENTATION MICROSURGERY 281 or consequence of failure, and no further sur­ the break otherwise enlarged to permit inter­ gery was advised. nal drainage of SRF. Approximation of the pigment epithelium to the retina by deep Discussion indentation also permits endolaser photo­ A single small break can cause a total detach­ coagulation of the retina adjacent to the break ment of the retina while the retina will as an alternative to transcleral cryopexy. withstand the tangential tractional effects of Although the actual number of retinal widespread epiretinal membranes without breaks in the eyes in this series is unknown, a significant elevation provided no breaks are break or multiple breaks were successfully created or re-opened. It follows that the identified in 95% of the procedures. How­ methodology of internal searching for breaks ever, internal searching is not without its ranks at least equal in importance to sophisti­ dangers. Lens damage, retinal incarceration cated membrane dissection techniques in in entry sites and subretinal infusion are risks closed microsurgery for rhegmatogenous inherent in any closed microsurgical approach detachment. Posterior breaks are readily to a bullous retinal detachment while choroi­ identified by standard methods of endo­ dal haemorrhage is an ever-present concern in illumination combined with aspiration; the highly myopic or inflammed eyes. Further­ schlieren of relatively dense SRF entering a more, the possible introduction of new pre-retinal compartment filled with infusion vitreoretinal tractional problems owing to fluidconfirms the full-thickness retinal discon­ either vitreous incarceration in entry sites6 or tinuity. However, the majority of retinal continuing hypocellular gel contraction in the breaks are located peripherally, especially at vitreous base (anterior loop traction)7 under­ the posterior border of the vitreous base, and scores the need for indentation vitrectomy of special techniques are required for visualisa­ basal gel and consideration of encirclement in tion of this area during vitrectomy. Prismatic these cases. Nevertheless, the newly-dis­ fundus-viewing lenses can be helpful for this covered breaks can be closed precisely (by purpose (and indeed the 3-mirror contact lens focal scleral buckling and/or retinal coagula­ is employed in some European centres even tion) and the associated retinal detachment for conventional retinal reattachment pro­ treated definitively. We therefore commend cedures). Alternatively, some surgeons tem­ the internal search for eyes with unseen reti­ porarily plug the sclerotomies after nal breaks, especially those which have pre­ vitrectomy and view the ocular interior with viously undergone failed retinal reattachment the indirect ophthalmoscope (Michels, per­ surgery. sonal communication). We prefer the tech­ nique described here whereby reflex-free References endoillumination is combined with surgeon­ 1 Gonin J: Treatment of detached retina by sealing the retinal tears. Arch Ophthalmol1930, 4: 621-5. controlled deep kinetic scleral indentation 2 Wong D, Billington BM, Chignell AH: Pars plana thus enabling the whole circumference of the vitrectomy for retinal detachments with unseen peripheral retina to be viewed up to and retinal holes. Graefe's Arch Clin Exp Ophthalmol including the ora serrata even in phakic eyes. 1987,225: 269-71. 3 Lincoff H and Geiser R: Finding the retinal holes. Breaks can be seen directly or the operculum Arch Ophthalmol1971, 85: 565-70. viewed in profile; small breaks are often best 4 Crick MDP and Chignell AH: Treatment of rheg­ seen on the slope rather than the summit of matogenous detachment without apparent holes. indentation. The zoom magnificationand bin­ Trans Ophthalmol Soc UK 1977, 99: 272-4. ocular viewing of the operating microscope 5Jagger JD, Cooling RJ, Fison LG, Leaver PK McLeod D: Management of retinal detachment serve to maximise the surgeon's capability to following congenital cataract surgery. Trans identify breaks. Where residual opacities in Ophthalmol Soc UK 1983,103: 103-7. the vitreous base preclude a complete periph­ 6 McLeod D: Fresh retinal detachments-the role of eral search, further indentation vitrectomy scleral buckling. Trans Ophthalmol Soc UK 1986, with the suction-cutter and coaxial micro­ 105: 480-8. 7 Charles S: Vitreous microsurgery. Second Edition scope illumination can be used and, if neces­ 1987. Pub!. Williams and Wilkins. Baltimore/ sary, the flapof any tear can be amputated or London.