Australian and New Zealmd Journal of Opbtbalmology ( 1 997) 25, 25-30

Original Article

Retinal detachment caused by retinal dialysis CJ Kennedy, FRACO, CE Parker, MD, and IL McAllister, FRACO Lions Eye Institute, Nedlands, Western Australia, Australia

ABSTRACT

Purpose: To determine the characteristics o f patients devel- oping retinal detachment secondary to retinal dialysis in Western Australia and t o confirm the clinical impression that these patients had a low rate o f proliferative vitreo- ret ino pathy (PVR) .

Methods: A retrospective analysis of the records of I60 I consecutive patients with rhegmatogenous retinal detach- ment identified 7 I patients in whom the retinal detachment was caused by a retinal dialysis.

Results: The majority o f these patients were young adults (mean age of 30 years) and the male to female ratio was I .3 : I . Seventy per cent of patients provided a history of significant trauma to the affected eye. Sporting injuries, assault, and motor vehicle injuries together accounted for 72% o f identifiable trauma. Examination revealed a dialysis of the inferotemporal quadrant in 75% of cases and despite obvious signs of chronicity of the associated retinal detach- ment (such as intraretinal macrocysts and demarcation lines) in approximately one-third of the eyes, only 5.6% developed grade C I PVR either pre- or postoperatively.

Conclusion: The present study supports the view that it is the low rate of PVR that explains the good prognosis and high surgical success rate for retinal detachments caused by retinal dialysis. It is postulated that a major reason for the low rate of PVR is that the vitreous base attachment t o the posterior margin of a retinal dialysis acts as a significant barrier t o the migration of potentially proliferative retinal pigment epithelial cells.This may lead to containment of the responsible proliferative cells within the loculated subretinal space.

Keywords: Posterior vitreous detachment, proliferative vit- reoretinopathy, retinal dialysis, vitreous base.

Retinal dialysis is the most common cause of retinal detach- ment in children and young adults.’ The term ‘retinal dialy- sis’ refers to a tear that i s located parallel to and in the immediate vicinity of the ora serrata. Most tears are within the retina; however, some are actually located in the non- pigmented ciliary epithelium just anterior to the ora.2 The vitreous base straddles the ora serrata and measures 3.3 mm nasally and 3.2 mm temporally in its anteroposterior dimen- sion.3 Although somewhat controversial, the factor that may perhaps best define a retinal dialysis, and explain its differ- ence in behaviour to a giant retinal tear, is that the break is located anterior to the posterior limit of the vitreous base insertion.2.M The persistence of the vitreous base attach- ment to the free margin of the detached retina at the poste- rior edge of the dialysis prevents the retina from folding back on itself, as often happens in giant retinal tears.? In the latter case, the retinal break is entirely posterior to the vit- reous base and the free margin of retina is not supported.‘ Posterior vitreous detachment, a prerequisite of other forms of rhegmatogenous retinal detachment, rarely accompanies detachments caused by retinal dialysis.2

Blunt trauma i s the most common identifiable cause of retinal dialysis and subsequent retinal detachment. The pathogenetic mechanisms involved have been studied in experiments that subjected porcine eyes to trauma.’ Anteroposterior compression of the globe causes equatorial expansion of the sclera. Expansion is maximal in the region of the vitreous base where the vitreous is firmly adherent to the non-pigmented ciliary epithelium and peripheral retina. Because the vitreous base is more rigid and less expansile than the sclera, significant vitreous traction forces are con- centrated in the region of the retina abutting the ora serrata. There is evidence that the quadrantic location of a retinal dialysis is affected by the location of the inciting trauma. Weidenthal and Schepens found that central corneal inden- tation caused nasal dialyses while projectile impaction on

W Correspondence: Dr IL McAllister, Lions Eye Institute, 2 Verdun Street, Nedlands, WA 6009, Australia

26 Kennedy et al.

the temporal limbus increased the incidence of temporal dialyses.' Penetrating trauma and ocular surgery are less fre- quent causes of retinal dialysis.*,Y

Non-traumatic retinal dialysis is a rarer and more contro- versial entity4t5r10 Although some still believe that all retinal dialyses result from blunt ocular trauma," there is growing evidence that not all dialyses are traumatic in origin. Lattice degeneration has been implicated histopath~logically,~ and Kinyoun and Knobloch have described non-traumatic dialy- ses associated with marked peripheral microcystoid changes and yellow-white vitreous opacities.4 Presumed non-traumatic dialyses mainly occur in the inferotemporal quadrant and are more commonly bilateral.5,'2 Genetic factors have also been implicated and this proposition has been supported by reports of familial clustering of cases of retinal dialysis.I3,l4 Although it is thought that even mild ocular trauma may precipitate retinal dialysis in predisposed individuals, there is no evidence that indirect trauma to the head in the absence of trauma to the globe is a significant cause of retinal dialysis.I5,l6

While traumatic retinal dialyses date from the time of injury,l7 the delay in presentation is often lengthy and may be measured in years. This is frequently a result of the slow rate of progression of retinal detachments associated with retinal dialysis18 and frequent paucity of symptoms until the pos- terior pole becomes involved by the detachment. A signifi- cant number of cases are therefore diagnosed incidentally. While long delays in presentation inevitably impair recall of previous potentially significant ocular injuries, the remem- brance of ocular injuries in the distant past, perhaps even decades previously, must be of questionable significance. The true proportion of retinal dialyses that are traumatic in origin has therefore been very difficult to determine with accuracy.

Retinal dialysis-related retinal detachments have a more favourable prognosis than other forms of rhegmatogenous detachment retinal detachment,lJ,'Y but the reason for this is not very clear. We believe that an inherently low tendency for development of proliferative vitreoretinopathy in these cases may be a prime factor. Few studies have previously addressed this.

The aims of the present study were therefore to deter- mine the characteristics of patients developing retinal detachment secondary to retinal dialysis in Western Australia, to determine their clinical course, and to deter- mine the incidence of proliferative vitreoretinopathy (PVR).

METHODS Cases of retinal detachment caused by a retinal dialysis in the absence of penetrating globe trauma were drawn from a review of the records of 160 1 consecutive patients with rheg- matogenous retinal detachment who had been seen over the past 17 years at the three hospital centres that comprised all

of the tertiary referral centres for vitreoretinal surgery in the state of Western Australia. A retrospective analysis was per- formed on the records of the 71 patients identified to have had a retinal detachment caused by retinal dialysis.

Data gathered from the records included documentation of a history of trauma, the interval from trauma to presenta- tion, symptoms at presentation, location and size of the dialysis, associated vitreoretinal findings, management, and outcome. Follow-up ranged from 1 week to 16 years with a mean of 22 months; 67% of patients were followed for more than 3 months.

RESULTS Of the 71 patients with retinal detachment caused by a retinal dialysis, 40 (56%) were male and 31 (44%) were female, resulting in a male to female ratio of 1.3 : I . The mean age of the patients at the time of diagnosis was 30 years with a range of 1 1 to 68 years; 63% of patients were 30 years of age or younger. The right eye was involved in 37 patients (52%) and the left eye in 3 3 patients (47%). One patient ( 1 %) had bilateral retinal detachments due to dialyses.

A positive history of blunt trauma was recorded for 50 (70%) of the cases and 60% of these were males. In the sub- group of patients who gave a history of trauma, almost one- third (32%) reported a sports-related accident. Squash, cricket and basketball were the most commonly cited sports. Assaults (28%) were the second most common mechanism by which ocular trauma was inflicted and these usually involved fist blows to the eye. A significant proportion of injuries (20%) were accidentally self-inflicted, often during household chores. Motor vehicle accidents (12%) were a less frequent cause of globe trauma.

The time period between the documented episode of trauma and the time of diagnosis of the detachment ranged from 1 day to an isolated extreme of 56 years. In 10% of patients this period was 1 week or less, in 17% it was between 1 week and 1 month, in 33% it was between 1 month and 1 year, in 15% it was between 1 and 5 years, and in 25% it was more than 5 years. All four patients with superonasal retinal dialyses presented within 1 year of the documented episode of trauma.

Symptoms prompting ophthalmologic evaluation were present in 52 patients (73%); 15 patients (21%) were asymp- tomatic and no data regarding symptomatology was avail- able on four patients (6%). The most common symptom, present in 59% of all patients, was blurred vision. The macula was detached in 88% of patients who presented with this complaint. Visual field loss was the main presenting symptom of 13% of all patients. Only 1 % of patients pre- sented with floaters as an isolated symptom. Interestingly, no patients reported photopsia.

Retinal detachment due to retinal dialysis 27

The most frequent location of the retinal dialysis was in the inferotemporal quadrant (Fig. 1). The inferotemporal quadrant was involved in 82% of cases and constituted the only involved quadrant in 68%. The superotemporal quad- rant was involved alone in 7%, and both temporal quadrants were involved in an additional 13%. The superonasal quad- rant was involved alone in 7%, the inferonasal quadrant alone in 3%, and both nasal quadrants in 1%. The final 1 % of the dialyses straddled the two inferior quadrants. In three (4%) of the 71 patients with retinal detachment due to retinal dialysis in one eye, the fellow eye harboured a retinal dialysis without retinal detachment.

In the vast majority of cases (90%) in which there was n o documented history of ocular trauma, and in most patients with a history of trauma in the distant past, the retinal dial- ysis involved the inferotemporal quadrant. All patients with a dialysis involving the superonasal quadrant had a docu- mented episode of trauma. The average size of retinal dialy- sis was a fraction under 2-clock-hours. There were n o apparent differences in the average size of the dialysis between the different quadrants or with regard to the pres- ence or absence of a documented preceding episode of trauma.

Associated vitreoretinal signs included demarcation lines in 19 patients (27%), intraretinal macrocysts in 6 patients (8%) and subretinal bands in 6 patients (8%). Four of the six patients with subretinal bands had no preretinal membrane formation. Marked peripheral cystoid degeneration was present in six patients (8%) and half of these had no docu- mented history of trauma. No patients were found to have horseshoe tears or other forms of retinal break in addition to the dialysis. Proliferative vitreoretinopathy (PVR) of grade C1 or greater was found in three patients (4%) at presenta- tion. The first of these patients had been hit with a crowbar 4 months prior to presentation and was noted to have a 30 degree superonasal dialysis with grade C 1 PVR. The second patient had been aphakic from infancy, following intracap- sular cataract extractions for congenital cataracts. He was intellectually disabled, had a history of alcohol abuse, and had suffered at least one blow to the eye from a clenched fist

Su perotem poral mSuperonasa'

I n fe r ot ern p o 4-1 n f e r o n asa I Figure 1. Quadrantic location of retinal dialyses.

over 5 years prior to presentation. This patient had a 3- clock-hour temporal retinal dialysis with grade C2 PVR. The duration of macular detachment could not be deter- mined because the patient wore n o aphakic correction and had not noted a change in vision. The third patient had a 2- clock-hour superotemporal dialysis with grade D 1 PVR. This patient had Down syndrome and was also significantly intellectually disabled. The history of preceding ocular trauma was uncertain.

The manner in which patients were managed depended upon the extent, duration and perceived stability of the retinal detachment. Management was interventional in 96% of cases. Three (4%) of the patients with longstanding, localized retinal detachments and prominent demarcation lines were followed cIosely by observation alone. One patient (1 %) had laser treatment and four patients (6%) had cryopexy to demarcate a localized detachment. Scleral buckles were used to treat 62 patients (86%). The majority (84%) of these were applied in conjunction with 360' encir- cling bands. Two patients (3%) required vitrectomy and retinal tamponade with intravitreal C,F, gas in one case and silicone oil in the other.

Of the 64 patients who underwent surgery, 62 patients (97%) had their retina successfully re-attached 1 month fol- lowing a single procedure. Fifty-nine (92%) remained primary surgical successes for the duration of follow-up. Primary repair failed to re-attach the retina in two patients (3%) and the retinas of an additional three patients (4%) subsequently redetached 5 months, 2 years, and 6 years postoperatively. The first two of these retinal redetachments occurred as a result of reopening of the original dialyses, the former due to migration of a non-encircled buckling element and the latter due to PVR-related tractional opening of a previously closed dialysis. The third patient's retina rede- tached following the formation of a giant retinal tear adja- cent to the closed dialysis. All the preceding cases of primary and secondary failure were successfully managed with a second surgical procedure.

Pre-operative Snellen visual acuity of the patients who underwent surgery was 6/12 or better in 7%, 6/15-6136 in 21%, and 6/60 or less in 72%. Postoperative visual acuity, noted at the most recent examination, was 6/12 or better in 40%, 6115-6/36 in 30%, and 6/60 or less in 30% of all patients. The macula was detached in 50 (78%) of the patients who underwent surgery and in this group, postop- erative visual acuity improved in 64% of patients, was unal- tered in 32%, and declined in 4%. Cystoid macular oedema and lamellar macular hole formation were responsible for the latter cases in which visual acuity declined. Vision declined in four (29%) of the 14 patients who had surgery for retinal detachments that did not involve the macula. Causes for reduced postoperative vision in this group

28 Kennedy et al.

included cystoid macular oedema, premacular fibrosis (which subsequently required vitrectomy and membrane peeling), and an iatrogenic macular tear during vitrectomy. In one patient the cause was unspecified.

DISCUSSION Although the male to female ratio of 1.3 : 1 in our series is slightly lower than the ratios of 1.4 : 1 to 4.8 : 1 reported in previous series,',2,1l,l2 the age distribution of cases was similar. Whereas Hagler reported that only 36-45% of his cases of retinal detachment associated with retinal dialysis had suffered trauma,'J the 70% prevalence of documented ocular trauma preceding the diagnosis of retinal detachment in our study i s similar to the rates reported in studies by Ross (68%)" and Zion and Burton (66%).12 I t is believed by many that the greater incidence of ocular trauma in males accounts for the predominance of male cases.

The predominance of inferotemporal quadrant involve- ment in retinal dialysis has also been noted in previous studies.'.l2,18 The structure of the bony orbit and the pres- ence of the nose leaves the inferotemporal quadrant of the globe most vulnerable to trauma and Bell's phenomenon may increase the exposure of this quadrant.7," Microcystoid degeneration of this portion of the retina may also be a pre- disposing factor.4

Two important indicators of the chronicity of the retinal detachments in the present study are intraretinal macrocysts and pigmented subretinal demarcation lines related to retinal pigment epithelial (RPE) cell hyperplasia.',2,'','2.20.2' lntraretinal macrocysts were found in 7% of our cases and have been reported in 9-25% of cases in previous series of dialysis-related retinal detachment. 1,2,22,23 These cysts are much less common in other forms of rhegmatogenous retinal detachment where the reported incidence is 1.6-3%.2,22 The 27% incidence of demarcation lines docu- mented in our series is similar to the 25-55% incidence reported in other series. 1 Non-dialysis-related rhegmatoge- nous retinal detachments have a lower (10-1 8%) reported incidence of demarcation lines.2.22

The incidence of significant PVR reported in the present study was relatively low. Although few previous studies of retinal dialysis-related detachments have specifically exam- ined this, two previous reports are consistent with this finding.2,I' In Hagler's series, the incidence of grade C1 or greater PVR at presentation was reported to be 1.2% in cases with no history of trauma and 6.3% in cases that had suffered prior trauma.2 Postoperatively in that study, massive preretinal proliferation developed in 0.9% of eyes without a history of trauma and in 2.5% of eyes where there was known preceding trauma. In another series, no patients were

icant PVR post-operatively." Other forms of rhegmatoge- nous retinal detachment have been noted to be associated with a significantly higher rate of PVR. In Hagler's study of 4837 eyes with rhegmatogenous retinal detachment, 19% of cases demonstrated grade C1 or greater PVR pre-operatively and 4.5% of cases developed massive preretinal proliferation postoperatively.2

We postulate that the low incidence of PVR associated with retinal dialysis is primarily due to a vitreous barrier to the liberation of RPE cells from the subretinal space. RPE cells are known to proliferate and form preretinal mem- branes along with glial and inflammatory cells.2427 Retinal detachment secondary to horseshoe retinal tears is always associated with posterior vitreous detachment.3 In such cases, the tear lies posterior to the vitreous base and a direct pathway exists from the subretinal space to the liquefied vit- reous behind the anteriorly collapsed posterior hyaloid face (Fig. 2). These retinal detachments are often bullous and tend to be much more mobile than retinal dialysis-related detachments. Eye movements cause eddy currents in the liquefied vitreous that increase the mobility of the detached retina and this may result in propulsion of loosened RPE cells from the subretinal fluid into the vitreous cavity.3 Disruption of the retinal surface by posterior vitreous detachment may also liberate the retinal glial cells that par- ticipate in PVR.26J7

In contrast, the vitreous base in retinal dialysis-related detachments remains adherent to the posterior margin of the detached retina (Fig. 3).2 The condensed collagen fibrils

Figure 2. Retinal detachment due to a retinal tear associated with posterior vitreous detachment. (a) Vitreous base anterior to tear, (b) retinal tear, (c) liberated RPE cells passing through retinal break

noted to have PVR preoperatively and 2% developed signif- to vitreous cavity, (d) posterior vitreous detachment

Retinal detachment due to retinal dialysis 29

reasons outlined previously, vitrectomy may have significant theoretical disadvantages. By removing the consensed vitre- ous base attachment to the posterior margin of the detached retina, the barrier to potentially proliferative cells may be removed and it is possible that by doing this, the retinal dial- ysis is effectively converted into a form of giant retinal tear with the attendant increased risk of PVR and tractional retinal redetachment. A variety of surgical procedures have been proven to be successful in retinal detachment associ- ated with retinal dialysis. The surgical technique preferred in our clinics was that of a localized explant covering the dialysis with an encircling band. Because these types of detachment tend to occur in young people, an encircling band was used to stabilize the explant and prevent late migration. Specific care was taken to ensure only minimal tension was applied to the band to minimize induced myopia.

It is unlikely that the low rate of PVR reported in the present retrospective study is simply a result of bias. The fact that the source of the study population was the sum of all tertiary referral centres for vitreoretinal surgery covering the geographically isolated State of Western Australia would suggest that any bias introduced should be toward cases with poorer prognosis. Several ophthalmologists in this State choose to refer only the more complicated retinal detachments to the referral centres, electing to manage the remainder themselves. This may explain why, at our study centres, retinal dialysis accounted for only 4.4% of the total number of retinal detachments seen. Other series have found 8-1 0% of retinal detachments to have been caused by retinal dialysis.'.2.".'2.23 The high success rate reported for retinal detachment caused by retinal dialysis coupled with the low rate of PVR and need for vitrectomy undoubtedly encourages more general ophthalmologists to manage these patients themselves. Although limited follow-up data was available in a number of our cases, this was virtually always due to the patients being returned for follow-up to their referring ophthalmologist or the ophthalmologist who ser- viced their sometimes geographically remote portion of the country. It is expected that late surgical failures or other sig- nificant complications would have been referred back to the tertiary referral centres.

In conclusion, in our study of patients with retinal detachment due to retinal dialysis we found there to be a low rate of associated PVR both pre- and postoperatively. This study supports the view that it is the low rate of PVR that explains the good prognosis and high surgical success rate for retinal detachments caused by retinal dialysis. We postulate that a major reason for the low rate of PVR is that the vitreous base attachment to the posterior margin of a retinal dialysis acts as a significant barrier to the migration of potentially proliferative RPE cells.

Figure 3 . Retinal detachment due to a retinal dialysis. (a) Vitreous base overlying dialysis, (b) retinal dialysis, (c) liberated RPE cells loculated in subretinal space, (d) posterior vitreous attached.

of the cortical vitreous in the region of the vitreous base may provide a significant barrier to the migration of RPE cells into the vitreous cavity and onto the inner retinal surface, thus leaving the RPE cells loculated within the subretinal space. A number of previous studies have commented on the pigmented granules seen in the region of the vitreous base in patients with dialysis.2,4 These may represent trapped ciliary epithelial cells,> avulsed cells from the neurosensory retina4 or RPE cells and may be testimony to the barrier function of the vitreous base. The vitreous base attachment to the margin of the detached retina in retinal dialysis may also provide a barrier in the opposite direction to liquefied vitre- ous entering the subretinal space and promoting the progress of retinal detachment. The lower incidence of pos- terior vitreous detachment noted in dialysis-related retinal detachments' would also reduce the incidence of liquefied vitreous abutting the margins of the retinal dialysis. The combination of the aforementioned factors may be respon- sible for the observed slower rate of progression of dialysis- related retinal detachments and higher incidence of subretinal bands and demarcation lines. The low frequency of posterior vitreous detachment and subsequent reduced mobility of the detached retina may also explain the absence of photopsia as a symptom in our patients.

Management of retinal detachments caused by retinal dialysis is highly successful in most cases1,2.19 and the present study supports the view that this is mainly attribut- able to the low rate of associated PVR. When surgery is nec- essary, vitrectomy is seldom required. Indeed, for the

30 Kennedy et al.

REFERENCES 1. Hagler WS, North AW. Retinal dialyses and retinal detach-

ment, Arcb. Opbtbalmol. 1968; 79: 376-88. 2. Hagler WS. Retinal dialysis: a statistical and genetic study to

determine pathogenic factors. Trans. Am. Opbtbalmol. SOC. 1980;

3. Lewis H, Kreiger AE. Rhegrnatogenous retinal detachment. In: Duane TD (ed.). Clinical Opbtbalmology, Vol. 3. Philadelphia: JB Lippincott, 1993; Ch. 27, 1-12.

4. Kinyoun JL, Knobloch WH. Idiopathic retinal dialysis. Retina

5. Smiddy WE, Green WR. Retinal dialysis: pathology and

6. Brockhurst RJ, Howard RO. Retinal dialysis. Arcb. Opbtbalmol.

7. Weidenthal DT, Schepens CL. Peripheral fundus changes associated with ocular contusion. Am. J. Opbtbalmol. 1966; 62:

8 . Litwin RL. Retinal dialysis secondary to use of the Kaufman vitrector. Am. J. Opbtbalmol. 1978; 85: 879-80.

9. Seelenfreund MH, Freilich DB. Retinal injuries associated with the cataract surgery. Am. J. Opbtbalmol. 1980; 89: 654-8.

10. Ross WH. Idiopathic retinal dialysis. Retina 1984; 4: 276-9. 1 1. Ross WH. Traumatic retinal dialyses. Arcb. Opbtbalmol. 198 1 ;

12. Zion VM, Burton TC. Retinal dialysis. Arcb. Opbtbalmol. 1980;

13 . Brown GC, Tasman WS. Familial retinal dialysis. Can. J. Opbtbalmol. 1980; t 5: 193-5.

14. Verdaguer TJ, Rojas B, Lechuga M. Genetical studies in non- traumatic retinal dialysis. Mod. Problems Opbtbalmol. 1975; 15: 34-9.

78: 687-733.

1984; 4: 9-14.

pathogenesis. Retina 1982; 2: 94-1 16.

1971; 85: 637-8.

465-77.

99: 1371-4.

98: 1971-4.

15. Benson WE, Shakin J, Sarin LK. Blunt trauma. In: Duane TD (ed.). Clinical Opbtbalmology, Val. 3 . Philadelphia: JS Lippincott, 1993; Ch. 31, 1-14.

16. Doden W, Stark N. Retina and vitreous findings after serious indirect trauma. Klin. Montatsbl. Augenbeilkd. 1974; 164: 32-40.

17. Tasman W. Peripheral retinal changes following blunt trauma. Trans. Am. Opbtbalmol. Soc. 1972; 70: 190-7.

1 8 . Johnston PB. Traumatic retinal detachment. Br. J. Opbtbalmol.

19. Howard RO, Gaasterland DE. Giant retinal dialysis and tear:

20. Cox MS, Schepens CL, Freeman HM. Retinal detachment

21. Marcus DF, Aaberg TM. lntraretinal macrocysts in retinal

22. Michels RC, Wilkinson CP, Rice TA. Retinal Detachment. St

23. Chignell AH. Retinal dialysis. Br. J. Opbtbalmol. 1973; 57:

24. Mandelcorn MS, Macherner R, Fineberg E, Hersch SB. Proliferation and metaplasia of intravitreal retinal pigment epithelial cell autotransplants. Am. J. Opbtbalmol. 1975; 80: 227-37.

25. Macherner R, Van Horn DL, Aaberg TM. Pigment epithelial proliferation in human retinal detachment with massive periretinal proliferation. Am. J. Opbtbalmol. 1978, 85: 181-91.

26. Van Horn DL, Aaberg TM, Machemer R, Fenzl R. Glial cell proliferation in human retinal detachment with massive periretinal proliferation. Am. J. Opbtba lmol . 1977; 84:

27. Grisanti S, Guidry C. Transdifferentiation of retinal pigment epithelial cells from epithelial to rnesenchymal phenotype. Invest. Opbtbalmol. Vis. Sci. 1995; 36: 391-405.

1991; 75: 18-21.

Surgical repair. Arcb. Opbtbalmol. 1970; 84: 312-15.

due to ocular contusion. Arcb. Opbtbalmol. 1966; 76: 678-85.

detachment. Arcb. Opbtbalmol. 1979; 97: 1273-5.

Louis: CV Mosby, 1990; 638-9.

572-7.

38 3-9 3 .