CRYOSURGERY FOR ACUTE RETINOPATHY OF PREMATURIT Y: FAC TORS ASSOCIATED WI TH TREATMENT SUCCESS AND FAILURE W. E. SCHULENBURG and J. F. ACHESON London
SUMMARY Cryosurgery for stage 3 plus acute retinopathy of pre maturity (ROP) increases the probability of disease regression, but up to 25% of eyes may progress to retinal detachment and blindness in spite of treatment. In a series of 37 eyes in 23 patients treated at the Hammer smith Hospital an overall 75% of eyes reached a favour able outcome. We present these results in detail and analyse the apparent causes.of treatment failure. Poor anatomical results (total retinal detachment and traction detachment involving the macula) were associ ated with inappropriate cryoprobe design and with zone 1 and rush-type disease. Poor functional results in the
presence of a largely flat retina (marked macular ectopia and high myopia) appeared to be associated with delayed treatment during the period of evolution of the stage 3 lesion, and with the appearance of a more highly differen tiated vascular shunt with early localised forward vit reous invasion. Suggestions for the refinement of existing gradings of acute ROP are made.
Acute retinopathy of prematurity (ROP) may affect up to 50% of infants at risk (born at less than 31 weeks ges tational age) but severe disease is rare and spontaneous regression without clinically significant permanent seque lae is the commonest outcome.I-3 The reported incidence of sight-threatening stage 3 disease depends on surveil lance criteria and ranges from 4-10% of the at-risk group?,4--6 In terms of ROP as a cause for blindness, an annual incidence of 550 cases in the USA and 50-100 cases in the UK are estimatedY Cryosurgery was first reported to induce regression of acute ROP by Yamashita in 1972.9-10 Since then, many conflicting reports regarding its efficacy have been pub lished but recently the authoritative American collabora tive Cryo-ROP study group4--6 has established both From Western Ophthalmic Hospital, London NWI and Hammer smith Hospital, London W 12. Correspondence to W. E. Schulenburg, FRCS, FCOphth, Western Ophthalmic Hospital, Marylebone Road, London NWI 5YE.
Eye (1992)6,215-220
improved anatomical and functional outcomes from appropriately timed treatment. In the Cryo-ROP study, eyes with advanced stage 3 disease affecting eight cumu lative or five contiguous clock hours in the presence of 'plus' features of posterior vascular congestion were said to have 'threshold' disease and without treatment were predicted to have a 50% chance of progression to total ret inal detachment or traction detachment involving the mac ula (referred to as 'an unfavourable anatomical outcome', corresponding to stages 5 and 4b in the International Classifications) and were randomised into treatment and non-treatment groups. Forty-three per cent of such eyes progressed to an unfavourable outcome at three months without treatment, and a significantly lower 21.8% with treatment. This represents a treatment failure rate, in purely anatomical terms, of about one eye in five. When the functional results were reported, a higher proportion of treated eyes had a poor outcome. This largely reflects the poor performance of eyes with traction detachments not involving the macular (stage 4a) in which macular ectopia may be severe. Although much progress has been made, we are still far from achieving our objectives of preventing blindness and the milder forms of visual disability in the very prematurely born infant. To achieve this goal, consideration must be given to case selection, the appropriateness of treating one eye or both eyes, method of treatment, timing of treatment, probe design and the extent of ablation required to induce regres sion of acute stage III ROP. These points are considered using our experience of cryotherapy over an eight year period in single unit. Patients and method
An ROP surveillance programme has been in operation at the Hammersmith Hospital since 1983. Infants were examined at weekly intervals from the age of three or four weeks onwards by the same clinician. The pupils were dilated using cyclopentolate 0.5% eye drops and phenyl ephrine 2.5% eye drops prior to examination by indirect ophthalmoscopy.
216
W. E. SCHULENBURG AND J. F. ACHESON Table la.
Group I: Adult 3 mm cryo probe: Paired eye study
Total patients 8, total eyes treated 8 Mean birthweight 870 g (standard deviation, s.d. 182) Mean gestational age at birth 26 weeks (s.d. 1.6) Mean age at treatment 40.75 weeks (s.d. 5.6) Mean follow-up time 44 months (s.d. 23)
treated in all new cases that followed. Zone I disease was excluded from the paired eye study and both eyes were treated to maximise possible benefit to the infant in view of the inherent poor prognosis of such cases.12,13 Three different probes were used. The 3 mm Amoils retinal cryo-probe was used during the paired eye study. The probe was rather bulky, not suitable for the small orbit of the premature infant and was replaced by a 1.5 mm Frigitron paediatric cryoprobe. Disappointing results led to the development of a 2x4 mm cryoprobe specifically designed for the orbito-ocular relationships of the pre mature infant (Fig. 1). Additional infants with threshold disease identified by other ophthalmologists in referring hospials were eval uated at their own neonatal unit and transferred to the Hammersmith Hospital for treatment within 48 hours if they qualified for treatment. All infants were followed through the outpatient depart ment two months after discharge and then at six monthly intervals. One set of twins with rush disease was lost to follow-up after 26 months. Anatomical success was defined as a flat posterior pole without a temporal retinal fold or retrolental membrane. Visual function was assessed independently by an orthoptist using the Catford drum or Sherridan-Gardiner single and linear letter testing as appropriate. An examination under anaesthesia was performed on at least one occasion in all infants older than 2 years for detailed retinal examination, measurement of intraocular pressure, refraction and photography. Acute and cicatricial ROP changes were documented according to the international classification systems in general use.14,15
Fig. 1 Cryo probes used (from left to right): 3 mm adult retinal probe, 2 x4 mm paediatric T-bar probe, 1.5 mm paediatric probe.
A prospective paired eye study was conducted between 1983 and 1988 to assess the value of cryotherapy in acute stage 3 disease. Treatment was applied to the right eye in symmetrical disease and to the more severely affected side in asymmetrical cases. The natural history was studied in the control eye and formed part of an earlier report.11 All infants with confirmed acute ROP remained in the unit until either spontaneous regression occurred or treatment was applied. Treatment was considered in eyes with a stage 3 shunt involving eight or more clock hours together with the progressive development of the plus features of posterior vascular congestion with or without additional iris congestion. These criteria were broadly equivalent to the definition of threshold disease used in the Cryo-ROP study. Informed consent was obtained and treatment was applied to one or both eyes in the neonatal unit under general anaesthesia by the same surgeon in all cases. Cryosurgery was directed at the nonvascularised retina up to, but not including the shunt and access was by radial conjunctival incision where the probe could not be intro duced sufficiently posteriorly. A single freeze thaw tech nique was employed in a scatter fashion. The paired eye study was terminated with the publication of the first Cryo-ROP study results in 1988 and both eyes were Table lb.
Group I: Adult 3 mm cryo probe: Paired eye study Final stage ROP Eye treated
1. 2. 3. 4. 5. 6. 7. 8.
Results
Treatment was applied in a single session and was com-
R L R R R R R L
Best corrected visual acuity
R
L
R
L
0
4a 5 4a 5 5 5 5
3/60 na 6/12 6/60
0 0
0 :5 0 0 0
Q
PL 6/18
§g1 6/6
Key to all tables: 6/6 = treated eye. * referred with plus disease by another ophthalmologist. R rush disease f&f = fixes and follows PL perception of light na = information not available.
Macular drag R
L
3/60
++
++
6/12 PL PL PL PL 6/6
+
+
++ + +
+
+
Refraction R
L
-13 -13 ilied at 12 months -1/50 -1.50 -9.0 -4.0 -6.0 -0.75
-0.5
CRYOSURGERY FOR ACUTE RETINOPATHY OF PREMATURITY Table IIa.
probe
Group II: Bilateral treatment in rush disease: adult 3 mm
Total patients 2, total eyes treated 4 Mean birthweight 740 g Mean gestational age at birth 25 weeks Mean age at treatment 32 weeks Mean follow-up time 32 months
Table lIb.
probe
R 5
o
Best corrected visual acuity
L
R
4a
lost to follow-up lost to follow-up
o
L
pleted in all infants without complication. None of the infants died during the postoperative period and all cases could be extubated within 24 hours of completion of treat ment without complications. A total of 23 infants were studied. The results are reported in four groups: (i) Paired eye study. (ii) Rush disease. (iii) Both eyes treated with 1.5 mm probe. (iv) Both eyes treated with 2x4 mm probe. Group I-Paired eye study
A 3 mm Amoils adult retinal cryoprobe was used and eight eyes in eight infants were treated. (For details see Table la.)
The results are summarised in Table Ib under the column headed 'final stage of ROP'. An anatomical success was achieved in six treated eyes, treatment failed in two eyes which progressed to stage 5. One of these had gross asym metrical disease with a persistent hyaloid artery on the treated more severely affected side, the importance of which remains unclear. The control eye regressed. Four control eyes progressed to stage 5, one to stage 4a and three eyes regressed spontaneously. Regression occurred in both the control and treated eye in one case and treat ment may not have influenced the final outcome. Iris con gestion and pupil rigidity was present in three eyes, two in Group III: Bilateral treatment: 1.5 mm paediatric probe
Four eyes in a pair of twins were treated. Both had rush disease (Table IIa). Anatomical and functional results
The twin with a birthweight of 680 g progressed to stage 5 in the right eye and 4a in the left with severe temporal drag. The second twin, with a birthweight of 800 g, regressed on both sides and had fixation and following when last seen before they were lost to follow up (Table lIb). Group III-Bilateral treatment: 1.5 mm paediatric
Four eyes in two patients were treated with a small tipped probe (Table IlIa). Anatomical and functional results
All four eyes progressed to stage 5 disease with perception of light only (Table IIIb). Group IV-Bilateral treatment: 2x4 mm ROP probe
Twenty-one eyes in 11 patients were treated. Case no 22 already had vitreous involvement in his right eye when first examined and treatment was only applied to the left eye (Table IVa).
Bilateral regression occurred in three infants, one eye regressed and the second eye progressed to stage 4a in three infants, one eye regressed and the other progressed to stage 5 (one not treated) in two infants, two progressed
Group III: Bilateral treatment: 1.5 mm paediatric probe Final stage ROP
11.* 12.*
Group II-Bilateral treatment: 3 mm probe
Anatomical results
Total patients 2, total eyes treated 4 Mean birthweight 915 g Mean gestational age at birth 26 weeks Mean age at treatment 37 weeks Mean follow-up time 26 months
Table Illb.
The functional results are summarised in Table lb. One infant died at the age of one year. In the remaining 14 eyes, nine had measurable visual function, five had 6/18 or bet ter, and all nine had myopic refractive errors. Those eyes with high myopia (-9.0 dioptres and -13.0 D) also had significant macular ectopia. In three infants the control eye regressed spontaneously. In two of these, there was no difference between the refractions of the treated and untreated eyes (-1.5 D right and left and -13.0 D right and left).
probe
Anatomical results
Table IlIa.
the treated group and one control eye, all of which pro gressed to stage 5 disease. The onset of stage 5 was delayed by two weeks on the treated side in the infant that progressed to bilateral stage 5 disease. Functional results
Group II: Bilateral treatment in rush disease: adult 3 mm
Final stage ROP
217
Best corrected visual acuity
R
L
R
L
5 5
5 5
PL PL
PL PL
Table IVa.
Group IV: Bilateral treatment: 2x4 mm T-bar ROP probe
Total patients 11, total eyes treated 21 Mean birthweight 940 g (standard deviation, s.d. 175) Mean gestational age at birth 25.5 weeks (s.d. 1.5) Mean age at treatment 37.5 weeks (s.d. 3.0) Mean follow-up time 14 months (s.d. 9)
218
W. E. SCHULENBURG AND 1. F. ACHESON
Table IVb.
Group IV: Bilateral treatment: 2x4 mm T-bar ROP probe Final stage ROP
13. 14.R 15. 16. 17. 18.* 19.* 20.* 21. 22. 23.
R
L
0 0 0 4a 0 5 0 4a 4a 5 4a
0 0 0 0 4a 0 4a 4a 5 0 4a
R 6/6 6/9 6/9 6/60 f&f PL 6/9 na na na na
L 6/6 6/9 6/9 6/36 f&f 6/18 6/60 na na na na
Refraction
Macular drag
Best corrected visual acuity R
L
+
+
++
++
+
+
.+
na na na na
na na na na
R
L
-0.5 -1.25 -1.0 -14.0 -10.0
-0.5 -1.25 -1.0 -12.0 -10.0 -1.75 -6.0 na na na na
-6.0 na na na na
Note: in case 22, left eye treated only.
to stage 4a in both eyes and in one infant, one eye reached stage 5 and the other stage 4a. Functional results
The three infants with bilateral full regression had the best recorded visual function. All three were low myopes, VA 6/6 BE (-0.5 DR & L), VA 6/9 BE (-1.00 D R & L) and VA 6/9 BE (-1.25 D R & L). In a further four cases with recorded visual function, one infant was blind in the right eye and achieved VA L 6/18 (-1.75 D), one had full regression in the left eye and stage 4a in the right eye with bilateral moderate macular'ectopia VA 6/60 R (-14.0 D) 6/36 L (-12.0 D), one infant had full regression in the right eye and stage 4a in the left with bilateral mild mac ular ectopia. At 18 months she could fix and follow and behaved like a seeing infant, but co-operation was poor and an accurate measurement of visual function failed. A further case had full regression on the right and stage 4a on the left side VA R 6/9 (-6.00 D) VA L 6/60 (-6.00 D), mild macular ectopia was present on the left side. The remaining four infants were too young for visual assessApproximate dimensions of cryo application
using 1.5=
probe tip
-:...::=:::;....--
ment. Cases 18 to 23 were transferred from outlying hos pitals. With the exception of case 19, cases 18-23 all had well differentiated fibrovascular membranes projecting into the vitreous when first seen. ROP 2x4 Cryoprobe
Smaller paediatric cryoprobes became available follow ing the preliminary Cryo-ROP report. We used a 1.5 nun paediatric probe in four eyes with threshold disease. The same selection criteria, observer status and treatment method was applied as in all the other infants. All fOUI eyes progressed to stage 5 cicatricial ROP. The only change in management was the different size cryoprore and was considered a possible cause for failure in these eyes. The retinal surface area involved in treatment of Zone II disease was calculated based on measurements by Sorsby and Sherridan16 (Figs 2 and 3, Table V). It was estimated that 72% of avascular retina needs to be ablated to induce regression and can be achieved by approximately 35 applications employing the 3 nun probe. In the case of the 1.5 mm probe, 144 applications were required to achieve the same degree of ablation reaching an unpractical level to achieve. A cryoprobe with a larger surface contact area and slender stem dimensions allows easier access to the neonatal orbit and appeared to make undertreatment less likely.
fl'fo--:-""",,,..-"""':==:"""'--���� Ora Semt.:
anterior Iintit of treatment area h=6mm
-'---t-\'"'-:....;.;..:..-:.."-'-�-'-...;��:...:.:.;��
Equator: Approximate posterior limit of treatment area
Approximate dimensions
of cryo application using 3= probe tip
Fig. . 2
�
chematic diagram illustrating increased number of . app/tcatlOns required to ablate avascular retina between shunt and ora with a smaller cryoprobe tip size.
Fig. 3 Schematic diagram illustrating calculation of the approximate dimensions of avascular retina for cryoblation.
CRYOSURGERY FOR ACUTE RETINOPATHY OF PREMATURITY Table V.
Comparison of alternative cryo-probe designs 3 mm probe
Estimated probe tip surface area 7 sq mm Area covered by 35 cryo applications 250 sq mm Per cent treatment area (345 sq mm) covered 72% Number of applications required to cover 72 per cent of treatment area 35 Area covered by 50 applications 350 sq mm
1.5 mm probe 1.75 sq mm 62 sq mm 18% 144 88 sq mm
Discussion
In this study the single observer status allowed maximum consistency in case selection, method of treatment and assessment of treatment response. This permitted the identification of differences in respOJ;1se with different probes, and of different stage 3 shunt morphologies. Some infants have been followed up to 72 months providing some information on the longer term results and effective ness or complications of cryotherapy. The favourable effect of cryotherapy was evident in anatomical and functional terms. No long-term complica tions were observed in any of the infants during the fol low-up period that could be attributed to cryotherapy. Functional evaluations were documented on 20 eyes. Thirteen out of 20 eyes achieved 6/18 or better and the worst result was 3/60 in both eyes of one infant. Three infants in the paired eye study underwent spontaneous regression in the control eye. Final visual function remained the same in both the treated and untreated eye in these cases and cryotherapy did not seem to have any influence on the final refractive error or functional result. In two of the three infants the refractive error was the same in both eyes (-13 DR & L and -1.5 DR & L) and in the third infant anisometropia existed with a refractive error of -0.75 D on the right untreated side and -0.50 D on the
219
treated side. High myopia (>-6 D) was associated with a poorer functional outcome, more severe retinal drag and macular ectopia, caused by more advanced retinal cicatri cial disease where morphologically the retinal lesion was better differentiated and projected deeper into the vitreous at the time of treatment. Cryotherapy has in the past been considered as a possible cause for high myopia.17-18 A more important factor may be timing of treatment with high myopia in those infants where treatment was longer delayed, further supporting the argument for early treat ment once stage 3 had been reached. If high myopia is associated with more advanced stage 3 disease reached at the time of regression (spontaneous or induced) similar untreated eyes will have a higher incidence of progression in which case abnormal ocular growth and myopia will be masked by retinal detachment and endstage disease in the control eye. A favourable outcome was achieved in 75% of treated eyes. If the 21 eyes treated with the 2x4 mm probe are considered, a favourable outcome was achieved in 86% of eyes. All eyes treated with the 1.5 mm probe failed and the high failure rate may be a result of the large number of applications required for adequate ablation with the small surface contact area of the probe. It underlines the import ance of achieving adequate ablation of the nonvascular retina which is not only dependent on technique but also probe design. The cryoprobes used in the Cryo-ROP study was not standardised and three different probes were used possibly explaining some of the unfavourable results in the treated eyes. Of eight infants transferred from outlying hospitals who had bilateral treatment, five had already progressed beyond what we morphologically would have regarded as the optimum timing for treatment on both sides and finally progressed to stage 4a in six and stage 5 in three eyes (one
Fig. 4A Early stage 3 acute ROP with an amorphous salmon
Fig. 4B More advanced stage 3 acute ROP with a well
coloured vascular shunt.
differentiated, forward projecting fibrovascular shunt.
220
W. E. SCHULENBURG AND 1. F. ACHESON
untreated). The retinal lesion had progressed from a rather amorphous salmon coloured shunt, typical for early stage III, to a more differentiated fibrovascular lesion which projected deeper into the vitreous (Figs 4A and B). This may result in localised vitreoretinopathyII leading to more advanced cicatricial disease with localised surface or transgel traction with or without traction retinal detachment. Cases 18 to 23 were transferred from other units and these infants did worse from an anatomical point of view with three stage 5 and six at stage 4a after regression. One eye already had vitreous involvement when first examined and was considered beyond treatment. With the narrow treatment window in the natural history of the disease any unnecessary delay in treatment should be prevented which can be used against an argument for centralisation of cryotherapy to selected treatment centres. Threshold disease as defined in the Cryo-ROP trial was chosen as the point at which without treatment the probability of retinal detachment becomes 50%. This allowed statistical vali dation of the results, but does not necessarily define the best way for selecting infants for treatment. Even if spon taneous regression occurs in stage 3, the functional results may still be poor and there is evidence that earlier bilateral treatment may be associated with higher functional and anatomical success rates.19 Treatment outcome may be influenced by several variables, i.e. zonal disease, stage and maturity of the lesion, timing of treatment and quality of treatment (adequate treatment, single session). With the risk of treatment failure and the low reported incidence of cryo-related complications, it seems unreasonable to with hold a second chance for sight in these infants by with holding treatment to the second eye. A need may exist for the further development of the International Classification of Retinopathy of Prematurity so that treatment criteria can be more precisely standardised. Stage 3 might be quantified in terms of the extent of the lesion (clock hours), the morphological maturity of the lesion (early amorphous salmon coloured lesion or late differentiated fibro-vascular lesion projecting further into the vitreous), zone 2 might be subdivided into posterior (poor prognosis perhaps requiring earlier treatment) and anterior zones, and the presence and advance of 'plus' features disease more fully evaluated. Finally, systemic factors also need to be weighted and form part of the criteria for cryotherapy, as continuing hypoxaemia may be associated with more severe disease20 requiring early treatment if practicable. More effective treatment may be achieved if these extended criteria are applied, improving both anatomical and functional results. T he authors would like to thank Mr. John Arnold for his assistance with the clinical photographs, and Mr. Doig Sim monds for the line drawings. Key words: Fibroplasia.
CryosurgerylRetinopathy
REFERENCES I. Flynn JT, Banacalari E, Bachynski BN et al.: Retinopathy of prematurity: diagnosis, severity and natural history. Oph thalmol 1987, 94: 620-9. 2. Ng YK, Shaw DE . Fielder AR. Levene MI: Epidemiology of Retinopathy of Prematurity. Lancet 1988. ii: 1235-8. 3. Acheson JF and Schulenburg WE: Surveillance for. Ret inopathy of Prematurity in Practice: Experience from one neonatal intensive care unit. Eye1991.5: 80-5.
4. Cryotherapy for Retinopathy of Prematurity Study Group Multicenter Trial of Cryotherapy for Retinopathy of Pre maturity: Preliminary results. Arch Ophthalmol1988.106:
471-9. 5. Cryotherapy for Retinopathy of Prematurity Study Group Multicenter Trial of Cryotherapy for Retinopathy of Pre maturity: Three-month outcome . Arch Ophthalmol 1990,
108: 195-204. 6. Cryotherapy for Retinopathy of Prematurity Study Group Multicenter Trial of Cryotherapy for Retinopathy of Pre maturity:
One-year outcome-Structure and Function.
Arch Ophthalmol1990,108: 1408-16. 7. Phelps DL: Vision loss due to retinopathy of prematurity. Lancet 1981, i: 606. 8. Schulenburg WE: Unpublished survey 1986. 9. Yamashita Y: Studies on retinopathy of prematurity; III Cryocautery for retinopathy of prematurity. Jpn J Clin Oph thalmol1972,26: 385-93. 10. Payne JW and Patz A: Treatment of acute proliferative retrolental fibroplasia. Trans Am Acad Ophthalmol Otolar
yngol1972.76: 1234-46. II. Schulenburg WE, Prendiville A, Ohri R: Natural history of retinopathy of prematurity. Br J Ophthalmol 1987. 71: 837--43. 12. McPherson AR, Hittner HM, Lemos R: Retinal detachment in young premature infants with acute retrolental fibro plasia: report of 2 new cases. Ophthalmology 1982. 89:
1160-69. 13. Nissenkorn I, Kremer I, Gilad E, Cohen S, Ben-Sira 1: Rush type retinopathy of prematurity: report of three cases. BrJ
Ophthalmol 1987,72: 559-62. 14. The International Committee for the Classification of Ret inopathy of Prematurity: An International Classification of Retinopathy of Prematurity. Br J Ophthalmol 1984, 68:
690-7. 15. The International Committee for the Classification of the Late Stages of Retinopathy of Prematurity: An international classification of ROP II: The classification of retinal detach ment. Arch Ophthalmol1987, 105: 906-12.
16. Sorsby A and Sherridan M: T he eye at birth: measurements of the principal diameters in 48 cadavers. J Ana t1960, 94:
192-5. 17. Gordon RA and Donzis PB: Myopia associated with ret inopathy of prematurity. Ophthalmology1986,93: 1593-8.
18. Seiberth V. Knorz MC, Trinkman R: Refractive errors after cryotherapy in retinopathy of prematurity. Ophthalmolog
ica1990,201: 5-8. 19. Nissenkorn I, Ben-Sira I, Kremer I, et al.: Eleven years' experience with retinopathy of prematurity: visual results and contribution of cryoablation. Br J Ophthalmol 1991,
75: 158-9.
20. Ng YK, Fielder AR, Levene MI, Trounce JQ, McLellen N: Are severe Retinopathy of Prematurity and severe Per
of
PrematuritylRetrolental
iventricular Leucomalacia both ischaemic insults? Br J
OphthalmoI1989,73: 111-14.
SUMMARY Cryosurgery for stage 3 plus acute retinopathy of pre maturity (ROP) increases the probability of disease regression, but up to 25% of eyes may progress to retinal detachment and blindness in spite of treatment. In a series of 37 eyes in 23 patients treated at the Hammer smith Hospital an overall 75% of eyes reached a favour able outcome. We present these results in detail and analyse the apparent causes.of treatment failure. Poor anatomical results (total retinal detachment and traction detachment involving the macula) were associ ated with inappropriate cryoprobe design and with zone 1 and rush-type disease. Poor functional results in the
presence of a largely flat retina (marked macular ectopia and high myopia) appeared to be associated with delayed treatment during the period of evolution of the stage 3 lesion, and with the appearance of a more highly differen tiated vascular shunt with early localised forward vit reous invasion. Suggestions for the refinement of existing gradings of acute ROP are made.
Acute retinopathy of prematurity (ROP) may affect up to 50% of infants at risk (born at less than 31 weeks ges tational age) but severe disease is rare and spontaneous regression without clinically significant permanent seque lae is the commonest outcome.I-3 The reported incidence of sight-threatening stage 3 disease depends on surveil lance criteria and ranges from 4-10% of the at-risk group?,4--6 In terms of ROP as a cause for blindness, an annual incidence of 550 cases in the USA and 50-100 cases in the UK are estimatedY Cryosurgery was first reported to induce regression of acute ROP by Yamashita in 1972.9-10 Since then, many conflicting reports regarding its efficacy have been pub lished but recently the authoritative American collabora tive Cryo-ROP study group4--6 has established both From Western Ophthalmic Hospital, London NWI and Hammer smith Hospital, London W 12. Correspondence to W. E. Schulenburg, FRCS, FCOphth, Western Ophthalmic Hospital, Marylebone Road, London NWI 5YE.
Eye (1992)6,215-220
improved anatomical and functional outcomes from appropriately timed treatment. In the Cryo-ROP study, eyes with advanced stage 3 disease affecting eight cumu lative or five contiguous clock hours in the presence of 'plus' features of posterior vascular congestion were said to have 'threshold' disease and without treatment were predicted to have a 50% chance of progression to total ret inal detachment or traction detachment involving the mac ula (referred to as 'an unfavourable anatomical outcome', corresponding to stages 5 and 4b in the International Classifications) and were randomised into treatment and non-treatment groups. Forty-three per cent of such eyes progressed to an unfavourable outcome at three months without treatment, and a significantly lower 21.8% with treatment. This represents a treatment failure rate, in purely anatomical terms, of about one eye in five. When the functional results were reported, a higher proportion of treated eyes had a poor outcome. This largely reflects the poor performance of eyes with traction detachments not involving the macular (stage 4a) in which macular ectopia may be severe. Although much progress has been made, we are still far from achieving our objectives of preventing blindness and the milder forms of visual disability in the very prematurely born infant. To achieve this goal, consideration must be given to case selection, the appropriateness of treating one eye or both eyes, method of treatment, timing of treatment, probe design and the extent of ablation required to induce regres sion of acute stage III ROP. These points are considered using our experience of cryotherapy over an eight year period in single unit. Patients and method
An ROP surveillance programme has been in operation at the Hammersmith Hospital since 1983. Infants were examined at weekly intervals from the age of three or four weeks onwards by the same clinician. The pupils were dilated using cyclopentolate 0.5% eye drops and phenyl ephrine 2.5% eye drops prior to examination by indirect ophthalmoscopy.
216
W. E. SCHULENBURG AND J. F. ACHESON Table la.
Group I: Adult 3 mm cryo probe: Paired eye study
Total patients 8, total eyes treated 8 Mean birthweight 870 g (standard deviation, s.d. 182) Mean gestational age at birth 26 weeks (s.d. 1.6) Mean age at treatment 40.75 weeks (s.d. 5.6) Mean follow-up time 44 months (s.d. 23)
treated in all new cases that followed. Zone I disease was excluded from the paired eye study and both eyes were treated to maximise possible benefit to the infant in view of the inherent poor prognosis of such cases.12,13 Three different probes were used. The 3 mm Amoils retinal cryo-probe was used during the paired eye study. The probe was rather bulky, not suitable for the small orbit of the premature infant and was replaced by a 1.5 mm Frigitron paediatric cryoprobe. Disappointing results led to the development of a 2x4 mm cryoprobe specifically designed for the orbito-ocular relationships of the pre mature infant (Fig. 1). Additional infants with threshold disease identified by other ophthalmologists in referring hospials were eval uated at their own neonatal unit and transferred to the Hammersmith Hospital for treatment within 48 hours if they qualified for treatment. All infants were followed through the outpatient depart ment two months after discharge and then at six monthly intervals. One set of twins with rush disease was lost to follow-up after 26 months. Anatomical success was defined as a flat posterior pole without a temporal retinal fold or retrolental membrane. Visual function was assessed independently by an orthoptist using the Catford drum or Sherridan-Gardiner single and linear letter testing as appropriate. An examination under anaesthesia was performed on at least one occasion in all infants older than 2 years for detailed retinal examination, measurement of intraocular pressure, refraction and photography. Acute and cicatricial ROP changes were documented according to the international classification systems in general use.14,15
Fig. 1 Cryo probes used (from left to right): 3 mm adult retinal probe, 2 x4 mm paediatric T-bar probe, 1.5 mm paediatric probe.
A prospective paired eye study was conducted between 1983 and 1988 to assess the value of cryotherapy in acute stage 3 disease. Treatment was applied to the right eye in symmetrical disease and to the more severely affected side in asymmetrical cases. The natural history was studied in the control eye and formed part of an earlier report.11 All infants with confirmed acute ROP remained in the unit until either spontaneous regression occurred or treatment was applied. Treatment was considered in eyes with a stage 3 shunt involving eight or more clock hours together with the progressive development of the plus features of posterior vascular congestion with or without additional iris congestion. These criteria were broadly equivalent to the definition of threshold disease used in the Cryo-ROP study. Informed consent was obtained and treatment was applied to one or both eyes in the neonatal unit under general anaesthesia by the same surgeon in all cases. Cryosurgery was directed at the nonvascularised retina up to, but not including the shunt and access was by radial conjunctival incision where the probe could not be intro duced sufficiently posteriorly. A single freeze thaw tech nique was employed in a scatter fashion. The paired eye study was terminated with the publication of the first Cryo-ROP study results in 1988 and both eyes were Table lb.
Group I: Adult 3 mm cryo probe: Paired eye study Final stage ROP Eye treated
1. 2. 3. 4. 5. 6. 7. 8.
Results
Treatment was applied in a single session and was com-
R L R R R R R L
Best corrected visual acuity
R
L
R
L
0
4a 5 4a 5 5 5 5
3/60 na 6/12 6/60
0 0
0 :5 0 0 0
Q
PL 6/18
§g1 6/6
Key to all tables: 6/6 = treated eye. * referred with plus disease by another ophthalmologist. R rush disease f&f = fixes and follows PL perception of light na = information not available.
Macular drag R
L
3/60
++
++
6/12 PL PL PL PL 6/6
+
+
++ + +
+
+
Refraction R
L
-13 -13 ilied at 12 months -1/50 -1.50 -9.0 -4.0 -6.0 -0.75
-0.5
CRYOSURGERY FOR ACUTE RETINOPATHY OF PREMATURITY Table IIa.
probe
Group II: Bilateral treatment in rush disease: adult 3 mm
Total patients 2, total eyes treated 4 Mean birthweight 740 g Mean gestational age at birth 25 weeks Mean age at treatment 32 weeks Mean follow-up time 32 months
Table lIb.
probe
R 5
o
Best corrected visual acuity
L
R
4a
lost to follow-up lost to follow-up
o
L
pleted in all infants without complication. None of the infants died during the postoperative period and all cases could be extubated within 24 hours of completion of treat ment without complications. A total of 23 infants were studied. The results are reported in four groups: (i) Paired eye study. (ii) Rush disease. (iii) Both eyes treated with 1.5 mm probe. (iv) Both eyes treated with 2x4 mm probe. Group I-Paired eye study
A 3 mm Amoils adult retinal cryoprobe was used and eight eyes in eight infants were treated. (For details see Table la.)
The results are summarised in Table Ib under the column headed 'final stage of ROP'. An anatomical success was achieved in six treated eyes, treatment failed in two eyes which progressed to stage 5. One of these had gross asym metrical disease with a persistent hyaloid artery on the treated more severely affected side, the importance of which remains unclear. The control eye regressed. Four control eyes progressed to stage 5, one to stage 4a and three eyes regressed spontaneously. Regression occurred in both the control and treated eye in one case and treat ment may not have influenced the final outcome. Iris con gestion and pupil rigidity was present in three eyes, two in Group III: Bilateral treatment: 1.5 mm paediatric probe
Four eyes in a pair of twins were treated. Both had rush disease (Table IIa). Anatomical and functional results
The twin with a birthweight of 680 g progressed to stage 5 in the right eye and 4a in the left with severe temporal drag. The second twin, with a birthweight of 800 g, regressed on both sides and had fixation and following when last seen before they were lost to follow up (Table lIb). Group III-Bilateral treatment: 1.5 mm paediatric
Four eyes in two patients were treated with a small tipped probe (Table IlIa). Anatomical and functional results
All four eyes progressed to stage 5 disease with perception of light only (Table IIIb). Group IV-Bilateral treatment: 2x4 mm ROP probe
Twenty-one eyes in 11 patients were treated. Case no 22 already had vitreous involvement in his right eye when first examined and treatment was only applied to the left eye (Table IVa).
Bilateral regression occurred in three infants, one eye regressed and the second eye progressed to stage 4a in three infants, one eye regressed and the other progressed to stage 5 (one not treated) in two infants, two progressed
Group III: Bilateral treatment: 1.5 mm paediatric probe Final stage ROP
11.* 12.*
Group II-Bilateral treatment: 3 mm probe
Anatomical results
Total patients 2, total eyes treated 4 Mean birthweight 915 g Mean gestational age at birth 26 weeks Mean age at treatment 37 weeks Mean follow-up time 26 months
Table Illb.
The functional results are summarised in Table lb. One infant died at the age of one year. In the remaining 14 eyes, nine had measurable visual function, five had 6/18 or bet ter, and all nine had myopic refractive errors. Those eyes with high myopia (-9.0 dioptres and -13.0 D) also had significant macular ectopia. In three infants the control eye regressed spontaneously. In two of these, there was no difference between the refractions of the treated and untreated eyes (-1.5 D right and left and -13.0 D right and left).
probe
Anatomical results
Table IlIa.
the treated group and one control eye, all of which pro gressed to stage 5 disease. The onset of stage 5 was delayed by two weeks on the treated side in the infant that progressed to bilateral stage 5 disease. Functional results
Group II: Bilateral treatment in rush disease: adult 3 mm
Final stage ROP
217
Best corrected visual acuity
R
L
R
L
5 5
5 5
PL PL
PL PL
Table IVa.
Group IV: Bilateral treatment: 2x4 mm T-bar ROP probe
Total patients 11, total eyes treated 21 Mean birthweight 940 g (standard deviation, s.d. 175) Mean gestational age at birth 25.5 weeks (s.d. 1.5) Mean age at treatment 37.5 weeks (s.d. 3.0) Mean follow-up time 14 months (s.d. 9)
218
W. E. SCHULENBURG AND 1. F. ACHESON
Table IVb.
Group IV: Bilateral treatment: 2x4 mm T-bar ROP probe Final stage ROP
13. 14.R 15. 16. 17. 18.* 19.* 20.* 21. 22. 23.
R
L
0 0 0 4a 0 5 0 4a 4a 5 4a
0 0 0 0 4a 0 4a 4a 5 0 4a
R 6/6 6/9 6/9 6/60 f&f PL 6/9 na na na na
L 6/6 6/9 6/9 6/36 f&f 6/18 6/60 na na na na
Refraction
Macular drag
Best corrected visual acuity R
L
+
+
++
++
+
+
.+
na na na na
na na na na
R
L
-0.5 -1.25 -1.0 -14.0 -10.0
-0.5 -1.25 -1.0 -12.0 -10.0 -1.75 -6.0 na na na na
-6.0 na na na na
Note: in case 22, left eye treated only.
to stage 4a in both eyes and in one infant, one eye reached stage 5 and the other stage 4a. Functional results
The three infants with bilateral full regression had the best recorded visual function. All three were low myopes, VA 6/6 BE (-0.5 DR & L), VA 6/9 BE (-1.00 D R & L) and VA 6/9 BE (-1.25 D R & L). In a further four cases with recorded visual function, one infant was blind in the right eye and achieved VA L 6/18 (-1.75 D), one had full regression in the left eye and stage 4a in the right eye with bilateral moderate macular'ectopia VA 6/60 R (-14.0 D) 6/36 L (-12.0 D), one infant had full regression in the right eye and stage 4a in the left with bilateral mild mac ular ectopia. At 18 months she could fix and follow and behaved like a seeing infant, but co-operation was poor and an accurate measurement of visual function failed. A further case had full regression on the right and stage 4a on the left side VA R 6/9 (-6.00 D) VA L 6/60 (-6.00 D), mild macular ectopia was present on the left side. The remaining four infants were too young for visual assessApproximate dimensions of cryo application
using 1.5=
probe tip
-:...::=:::;....--
ment. Cases 18 to 23 were transferred from outlying hos pitals. With the exception of case 19, cases 18-23 all had well differentiated fibrovascular membranes projecting into the vitreous when first seen. ROP 2x4 Cryoprobe
Smaller paediatric cryoprobes became available follow ing the preliminary Cryo-ROP report. We used a 1.5 nun paediatric probe in four eyes with threshold disease. The same selection criteria, observer status and treatment method was applied as in all the other infants. All fOUI eyes progressed to stage 5 cicatricial ROP. The only change in management was the different size cryoprore and was considered a possible cause for failure in these eyes. The retinal surface area involved in treatment of Zone II disease was calculated based on measurements by Sorsby and Sherridan16 (Figs 2 and 3, Table V). It was estimated that 72% of avascular retina needs to be ablated to induce regression and can be achieved by approximately 35 applications employing the 3 nun probe. In the case of the 1.5 mm probe, 144 applications were required to achieve the same degree of ablation reaching an unpractical level to achieve. A cryoprobe with a larger surface contact area and slender stem dimensions allows easier access to the neonatal orbit and appeared to make undertreatment less likely.
fl'fo--:-""",,,..-"""':==:"""'--���� Ora Semt.:
anterior Iintit of treatment area h=6mm
-'---t-\'"'-:....;.;..:..-:.."-'-�-'-...;��:...:.:.;��
Equator: Approximate posterior limit of treatment area
Approximate dimensions
of cryo application using 3= probe tip
Fig. . 2
�
chematic diagram illustrating increased number of . app/tcatlOns required to ablate avascular retina between shunt and ora with a smaller cryoprobe tip size.
Fig. 3 Schematic diagram illustrating calculation of the approximate dimensions of avascular retina for cryoblation.
CRYOSURGERY FOR ACUTE RETINOPATHY OF PREMATURITY Table V.
Comparison of alternative cryo-probe designs 3 mm probe
Estimated probe tip surface area 7 sq mm Area covered by 35 cryo applications 250 sq mm Per cent treatment area (345 sq mm) covered 72% Number of applications required to cover 72 per cent of treatment area 35 Area covered by 50 applications 350 sq mm
1.5 mm probe 1.75 sq mm 62 sq mm 18% 144 88 sq mm
Discussion
In this study the single observer status allowed maximum consistency in case selection, method of treatment and assessment of treatment response. This permitted the identification of differences in respOJ;1se with different probes, and of different stage 3 shunt morphologies. Some infants have been followed up to 72 months providing some information on the longer term results and effective ness or complications of cryotherapy. The favourable effect of cryotherapy was evident in anatomical and functional terms. No long-term complica tions were observed in any of the infants during the fol low-up period that could be attributed to cryotherapy. Functional evaluations were documented on 20 eyes. Thirteen out of 20 eyes achieved 6/18 or better and the worst result was 3/60 in both eyes of one infant. Three infants in the paired eye study underwent spontaneous regression in the control eye. Final visual function remained the same in both the treated and untreated eye in these cases and cryotherapy did not seem to have any influence on the final refractive error or functional result. In two of the three infants the refractive error was the same in both eyes (-13 DR & L and -1.5 DR & L) and in the third infant anisometropia existed with a refractive error of -0.75 D on the right untreated side and -0.50 D on the
219
treated side. High myopia (>-6 D) was associated with a poorer functional outcome, more severe retinal drag and macular ectopia, caused by more advanced retinal cicatri cial disease where morphologically the retinal lesion was better differentiated and projected deeper into the vitreous at the time of treatment. Cryotherapy has in the past been considered as a possible cause for high myopia.17-18 A more important factor may be timing of treatment with high myopia in those infants where treatment was longer delayed, further supporting the argument for early treat ment once stage 3 had been reached. If high myopia is associated with more advanced stage 3 disease reached at the time of regression (spontaneous or induced) similar untreated eyes will have a higher incidence of progression in which case abnormal ocular growth and myopia will be masked by retinal detachment and endstage disease in the control eye. A favourable outcome was achieved in 75% of treated eyes. If the 21 eyes treated with the 2x4 mm probe are considered, a favourable outcome was achieved in 86% of eyes. All eyes treated with the 1.5 mm probe failed and the high failure rate may be a result of the large number of applications required for adequate ablation with the small surface contact area of the probe. It underlines the import ance of achieving adequate ablation of the nonvascular retina which is not only dependent on technique but also probe design. The cryoprobes used in the Cryo-ROP study was not standardised and three different probes were used possibly explaining some of the unfavourable results in the treated eyes. Of eight infants transferred from outlying hospitals who had bilateral treatment, five had already progressed beyond what we morphologically would have regarded as the optimum timing for treatment on both sides and finally progressed to stage 4a in six and stage 5 in three eyes (one
Fig. 4A Early stage 3 acute ROP with an amorphous salmon
Fig. 4B More advanced stage 3 acute ROP with a well
coloured vascular shunt.
differentiated, forward projecting fibrovascular shunt.
220
W. E. SCHULENBURG AND 1. F. ACHESON
untreated). The retinal lesion had progressed from a rather amorphous salmon coloured shunt, typical for early stage III, to a more differentiated fibrovascular lesion which projected deeper into the vitreous (Figs 4A and B). This may result in localised vitreoretinopathyII leading to more advanced cicatricial disease with localised surface or transgel traction with or without traction retinal detachment. Cases 18 to 23 were transferred from other units and these infants did worse from an anatomical point of view with three stage 5 and six at stage 4a after regression. One eye already had vitreous involvement when first examined and was considered beyond treatment. With the narrow treatment window in the natural history of the disease any unnecessary delay in treatment should be prevented which can be used against an argument for centralisation of cryotherapy to selected treatment centres. Threshold disease as defined in the Cryo-ROP trial was chosen as the point at which without treatment the probability of retinal detachment becomes 50%. This allowed statistical vali dation of the results, but does not necessarily define the best way for selecting infants for treatment. Even if spon taneous regression occurs in stage 3, the functional results may still be poor and there is evidence that earlier bilateral treatment may be associated with higher functional and anatomical success rates.19 Treatment outcome may be influenced by several variables, i.e. zonal disease, stage and maturity of the lesion, timing of treatment and quality of treatment (adequate treatment, single session). With the risk of treatment failure and the low reported incidence of cryo-related complications, it seems unreasonable to with hold a second chance for sight in these infants by with holding treatment to the second eye. A need may exist for the further development of the International Classification of Retinopathy of Prematurity so that treatment criteria can be more precisely standardised. Stage 3 might be quantified in terms of the extent of the lesion (clock hours), the morphological maturity of the lesion (early amorphous salmon coloured lesion or late differentiated fibro-vascular lesion projecting further into the vitreous), zone 2 might be subdivided into posterior (poor prognosis perhaps requiring earlier treatment) and anterior zones, and the presence and advance of 'plus' features disease more fully evaluated. Finally, systemic factors also need to be weighted and form part of the criteria for cryotherapy, as continuing hypoxaemia may be associated with more severe disease20 requiring early treatment if practicable. More effective treatment may be achieved if these extended criteria are applied, improving both anatomical and functional results. T he authors would like to thank Mr. John Arnold for his assistance with the clinical photographs, and Mr. Doig Sim monds for the line drawings. Key words: Fibroplasia.
CryosurgerylRetinopathy
REFERENCES I. Flynn JT, Banacalari E, Bachynski BN et al.: Retinopathy of prematurity: diagnosis, severity and natural history. Oph thalmol 1987, 94: 620-9. 2. Ng YK, Shaw DE . Fielder AR. Levene MI: Epidemiology of Retinopathy of Prematurity. Lancet 1988. ii: 1235-8. 3. Acheson JF and Schulenburg WE: Surveillance for. Ret inopathy of Prematurity in Practice: Experience from one neonatal intensive care unit. Eye1991.5: 80-5.
4. Cryotherapy for Retinopathy of Prematurity Study Group Multicenter Trial of Cryotherapy for Retinopathy of Pre maturity: Preliminary results. Arch Ophthalmol1988.106:
471-9. 5. Cryotherapy for Retinopathy of Prematurity Study Group Multicenter Trial of Cryotherapy for Retinopathy of Pre maturity: Three-month outcome . Arch Ophthalmol 1990,
108: 195-204. 6. Cryotherapy for Retinopathy of Prematurity Study Group Multicenter Trial of Cryotherapy for Retinopathy of Pre maturity:
One-year outcome-Structure and Function.
Arch Ophthalmol1990,108: 1408-16. 7. Phelps DL: Vision loss due to retinopathy of prematurity. Lancet 1981, i: 606. 8. Schulenburg WE: Unpublished survey 1986. 9. Yamashita Y: Studies on retinopathy of prematurity; III Cryocautery for retinopathy of prematurity. Jpn J Clin Oph thalmol1972,26: 385-93. 10. Payne JW and Patz A: Treatment of acute proliferative retrolental fibroplasia. Trans Am Acad Ophthalmol Otolar
yngol1972.76: 1234-46. II. Schulenburg WE, Prendiville A, Ohri R: Natural history of retinopathy of prematurity. Br J Ophthalmol 1987. 71: 837--43. 12. McPherson AR, Hittner HM, Lemos R: Retinal detachment in young premature infants with acute retrolental fibro plasia: report of 2 new cases. Ophthalmology 1982. 89:
1160-69. 13. Nissenkorn I, Kremer I, Gilad E, Cohen S, Ben-Sira 1: Rush type retinopathy of prematurity: report of three cases. BrJ
Ophthalmol 1987,72: 559-62. 14. The International Committee for the Classification of Ret inopathy of Prematurity: An International Classification of Retinopathy of Prematurity. Br J Ophthalmol 1984, 68:
690-7. 15. The International Committee for the Classification of the Late Stages of Retinopathy of Prematurity: An international classification of ROP II: The classification of retinal detach ment. Arch Ophthalmol1987, 105: 906-12.
16. Sorsby A and Sherridan M: T he eye at birth: measurements of the principal diameters in 48 cadavers. J Ana t1960, 94:
192-5. 17. Gordon RA and Donzis PB: Myopia associated with ret inopathy of prematurity. Ophthalmology1986,93: 1593-8.
18. Seiberth V. Knorz MC, Trinkman R: Refractive errors after cryotherapy in retinopathy of prematurity. Ophthalmolog
ica1990,201: 5-8. 19. Nissenkorn I, Ben-Sira I, Kremer I, et al.: Eleven years' experience with retinopathy of prematurity: visual results and contribution of cryoablation. Br J Ophthalmol 1991,
75: 158-9.
20. Ng YK, Fielder AR, Levene MI, Trounce JQ, McLellen N: Are severe Retinopathy of Prematurity and severe Per
of
PrematuritylRetrolental
iventricular Leucomalacia both ischaemic insults? Br J
OphthalmoI1989,73: 111-14.
