OF PREMATURITY: A PROSPECTIVE STUDY. REVIEW AT SIX MONTHS

2 2 2 2 D. LAWS1, D. E. SHAW ,3, J. ROBINSON ,3, H. S. JONES3, Y. K. NG , A. R. FIELDER ,3 Birmingham and Leicester

SUMMARY thalmic events. There have, however, been relatively few A cohort of infants of birth weight ::;;;1700 g studied pros­ studies on the ophthalmic outcome of low birthweight pectively for retinopathy of prematurity (ROP) has been infants which consider in detail neonatal data.lO-l9 reviewed at 6 months corrected age and the findings Recently we undertook a prospective study of ROP in related to the neonatal data. The overall incidence of stra­ the East Midlands of England which was specifically bismus was 6.4% (30/468), rising from 3.1 % (7/229) with­ designed to investigate its natural history.2o Here we pres­ out ROP to 29.2% (7/24) with stage 3. and ent the ophthalmic findings of this cohort obtained at 6 fusional ability were significantly related to presence and months corrected age and compare them with the findings severity of ROP, and abnormal neonatal cranial ultra­ obtained during the neonatal period. sound findings. Binocular visual acuity was measured in 340 infants between 20 and 40 weeks corrected age. Eight PATIENTS AND METHODS were subnormal, all due to neurological problems. For A prospective study to determine the incidence and sever­ the remainder, despite falling within the normal range, ity of ROP in a geographically defined population was there was a significanttrend (p<0.001)for lower acuities undertaken over a 23-month period between 1 July 1985 with increasing ROP severity. Cycloplegic refraction on and 31 May 1987 in the East Midlands of England.20 All 387 infants revealed, with increasing ROP severity: 1, a infants with a birthweight � 1700 g who survived 3 weeks significant trend towards ; 2, increased magni­ were enrolled from five neonatal units in the cities of tude of ; 3, alteration of the astigmatic axis; 4, Derby, Nottingham and Leicester. No infant received sup­ increased incidence of . For the first three plemental vitamin E. The epidemiological aspects of this categories there was an insignificant trend between no cohort have been reported?l ROP and stage 2, reaching significance only between Commencing at 3 weeks of age, ophthalmic examin­ stage 2 and 3. The predominant axis of astigmatism in infants without ROP was between 60° and 120°, but with ations were performed weekly during the infant's stay on ROP there was a significant trend away from this the neonatal unit until 12 weeks of age. After discharge, direction. infants were examined every 2 weeks until 12 weeks of age, and thereafter as clinically indicated, with a final Infants born preterm are at risk of experiencing a variety assessment at 6 months corrected age. All ophthalmic of ophthalmic problems during the neonatal period, some­ examinations at this time were performed by one ophthal­ times with sequelae which may become apparent later in mologist (ARF). Before 12 weeks, examinations were infancy, childhood or adult life. The best-known example performed following instillation of topical anaesthesia of the first is retinopathy of prematurity (ROP) and the (oxybuprocaine hydrochloride 0.4%) using an later associations or complications of pre term birth speculum and scleral indentation to visualise the entire include visual acuity deficits, ocular motility disorders, 360° of the retinal circumference. Indirect ophthalmos­ optic atrophy and refractive errors. \-9 In order to investi­ copy was performed using a 28 D preceded by dilata­ gate the causation of the ophthalmic sequelae of prematu­ tion of the with 0.5% drops ± rity, information on the perinatal period is mandatory, and 2.5%. Results were recorded on a computer ideally should include details of both systemic and oph- coding form with the acute changes of ROP documented

'Birmingham and Midland Eye Hospital, Birmingham, UK. according to the International Committee for Classifi­ 2University of Leicester Medical School, Leicester, UK, cation of Retinopathy of Prematurity.22 Cicatricial ROP 3University of Birmingham Medical School, Birmingham, UK. was recorded using the classification of Reese and Correspondence to: A. R, Fielder, MRCP, FRCS, FCOphth, Birmingham and Midland Eye Hospital, Church Street, Birmingham B3 associates.23 2NS, UK. The ophthalmic examination at around 6 months cor-

Eye (1992) 6,477-483 478 D. LAWS ET AL. rected age included visual acuity measurement, recorded Permission for this study was obtained from the Ethical using the acuity card procedure24 (by ARF except for Committees of Leicestershire, Nottingham and Southern orthoptic assessment by DP for some infants in one Derbyshire Health Authorities. centre). Binocular visions were measured; however, mon­ ocular acuities were also obtained in certain instances, RESULTS such as those infants with ROP stage 3 or above. Assess­ The cohort consisted of 572 infants who satisfiedthe birth­ ment of ocular motility included examination for the pres­ weight criteria. The gestational age (GA) and birthweight ence of strabismus and . Fusional ability was (Bwt) distributions of the cohort have been described pre­ assessed using a 20 0 base-out prism in front of each eye. viously.20 Mean GA was 30.8 weeks (SO 3.0 weeks) and was performed, with partjcular care being mean Bwt 1328 g (SO 270 g). There were 319 males and taken to remain on the optic axis, following 253 females ; 451 were singletons, liS twins and 6 the sur­ induced by instillation of cyclopentolate 1% eyedrops. vivors of multiple pregnancies. Ethnic origin was as fol­ Results were recorded with positive cylinder notation. lows: Caucasian, 465; Asian, 66; Afro-Caribbean, 24; Where given, mean spherical equivalents were calculated others, 17. Neonatal cranial ultrasound scan data were from the formula:25 available on 321 infants, and significant abnormal intra­ Spherical equivalent = [± sphere] + 0.5 [± cylinder] cranial events (see earlier) were noted in 85 (26.5%). Fundus examination at the 6 month visit was performed UJ1fortunately not all tests (e.g. cranial ultrasound) were using a 20 0 lens for greater magnification and a 28 0 performed on all infants. To avoid confusion, therefore, it lens to examine the retinal periphery. Scleral indentation is important that the reader does not always refer to the and an eyelid speculum were not used at this time. In addi­ entire cohort, but only to that proportion on whom that tion the degree of tortuosity of the retinal arterioles near particular investigation was performed. Direct compari­ the was recorded on a 5-point scale from 0 to son between tests is thus not always possible. ++++. Any other ophthalmic abnormalities, not directly The cohort was reduced to 524 by 3 months and oph­ related to ROP, were also noted. thalmic examinations were performed on 481 (91.8%) on Data were stored on the mainframe computer (IBM or after this corrected age. Data were collected on 3958 3090) at Birmingham University and analysed with the ophthalmic examinations,1213 performed after12 weeks. statistical package SPSS. Relationships between categor­ Acute ROP developed in 291 infants out of the 572 ical variables were investigated with the chi-squared test (50.9%). The maximum stage reached was I in 171 (Tables I-III, VII, VIII). Visual acuities and refractive (29.9%),2 in 93 (16.3%),3 in 25 (4.4%) and 4 in 2 (0.3%). components between ROP groups were tested by analysis Cicatricial ROP developed in 10 eyes of 6 (1.1 % of 572) of covariance with ROP group as factor in the model, and infants (ranging from 880 to 1300 g Bwt, and from 25 to age at testing as covariate. Various other factors that may 28 weeks GA), with grade 1 in 6 eyes, II in I eye, and III in confound the ROP group comparison were introduced as 3 eyes. The one infant who developed grade III cicatricial covariates in the model. In the case of overall significant disease bilaterally, died before the 6-month assessment, group differences, pairwise comparisons between groups and is not therefore included in the subsequent analysis. were conducted in order to identify the source of the One eye required cryotherapy, and this eye subsequently significance. developed cicatricial stage 3 ROP. There was a highly sig­ Non-ophthalmic data were also recorded. These nificant association between cranial ultrasound findings included details of pregnancy, labour, resuscitation, post­ and both the presence (p = 0.001) and stage (p<0.00I) of natal illness and cranial ultrasound (performed in four out ROP (Table I). of five neonatal units). Ultrasound abnormalities con­ The results of examination for strabismus were avail­ sidered significant were: able in 468 infants. Strabismus was present in 30 infants (6.4%): 20 convergent and 10 divergent (Table II). The I. Cystic periventricular leucomalacia or persistent flare. incidence of strabismus increased with both the presence 2. Ventricular dilatation persistent for more than I week. and stage of acute ROP as follows: no ROP, 7/229 (3.1 %); 3. Periventricular haemorrhage excluding subependymal stage I, 10/144 (6.9%); stage 2, 6/71 (8.5%), and stage haemorrhage. 3/4, 7/24 (29.2%). There was a significant relation 4. Miscellaneous neurological abnormalities (e.g. between strabismus and both presence and stage of ROP primary neurodevelopmental disorders). (p = 0.007 and p = 0.001 respectively), and abnormal

Table I. Cranial ultrasound findings and stage of acute ROP in 321 Table II. Stage of acute ROP and the presence of strabismus at 6 infants months corrected age for 468 infants

Stage of ROP Stage of ROP Convergent Divergent None Total Cranial ultrasound 0 2 3/4 Total 0 5 2 222 229 I 7 3 134 144 Nonnal 114 78 35 9 236 2 4 2 65 71 Abnonnal 23 26 23 13 85 3/4 4 3 17 24

Total 137 104 58 22 321 Total 20 10 438 468 ROP: SIX MONTH REVIEW 479

Table III. Stage of acute ROP and fusional ability at 6 months 80 454 corrected age for infants 6968 UI 59 Stage ofROP Positive Negative Total C 60 III 0 207 14 221 C 44 I 127 14 141 '0 40 2 57 12 69 a; 32 3/4 14 9 23 .Q 28 E � 405 49 454 20 Total Z

cranial ultrasound events in the neonatal period o (p

0 +1.45 +1.47 (-2.75 to +7.50) (-2.75 to 7.50) +1.53 +1.35 (-1.50 to +7.50) (-2.00 to +7.50)" 2 +1.06 +1.29 (-7.00 +3.50) (-9.50 +7.50)" 4�------r------.------.r------' to to o 2 3/4 3/4 -0.32 -0.42 (-3.25 to +1.50) (-5.00 to +2.75) ROP Stage With increasing ROP severity there is a trend towards myopia. from stage 0 to 2. reaching significance only between stage 2 and 3 (right eye, Fig. 1. Stage of acute ROP and mean visual acuity in cycles p

Table V. Stage of acuteROP and mean astigmatism (in dioptres) for Table VII. Stage of acuteROP and axis of astigmatism for right eye in 387 infants refracted between 20 and 45 weeks corrected age 387 infants refracted between 20 and 45 weeks corrected age. Numbers of infants for eachROP stage are given, with percentages in parentheses Right dioptres Left dioptres Stage of ROP (range) (range) Stage of ROP

0 +0.81 +0.80 Axis 0 2 3/4 Total (0.00 to +4.00) (0.00 to +4.50) +0.78 +0.99 60°-120° 88 48 19 I 156 (0.00 to +4.25) (0.00 to +4.00) (60.3%) (54.6%) (42.2%) (8.3%) 2 +0.99 +0.93 150°_30° 51 31 21 9 112 (0.00 to +2.75) (0.00 to +3.00) (34.9%) (35.2%) (46.7%) (75.0%) 3/4 +1.53 +1.60 Other 7 9 5 2 23 (0.00 to +3.75) (0.00 to +5.50) (4.80/c) (10.2%) (l1.I o/r) (16.7%) 146 88 45 12 291 With increasing ROP severity there is an increase in magnitude, only Total reaching significance between stages 2 and 3 (right eye. p = 0.0 I; left For infants who never developedROP the predominant axis is 60°_120°, eye. p = 0.(09). Full ranges in parentheses. but with increasing ROP severity there was a significant trend away from this axis (p = 0.0 I) and the proportion at other axes increased. increased significantly (left, p = 0.010; right, p = 0.002). The axis of astigmatism, expressed as a positive cylinder, neonatal period in order to study the natural history of by ROP stage for right and left eyes respectively is shown acute ROP.20 Here we have presented the ophthalmic find­ in Tables VII and VIII. Anisometropia � 1.0 D was pres­ ings of this cohort at 6 months corrected age and compared ent in 23 infants: 3.9% (7/181) of stage 0, 3.3% (4/123) of them with data obtained during the neonatal period. stage 1, 7.7% (5/65) of stage 2, and 41.2% (7/17) of stage 3 The incidence of strabismus (6.4%) was lower than that infants. These results show a significant relation noted in previous studies/-I 1.13-19 which ranged from 9% (p<0.001) between anisometropia and ROP stage, but the to 25%. The various cohorts are not directly comparable numbers at stage 3 were small. As only one eye of one and testing at 6 months almost certainly underestimated baby was treated with cryotherapy, no comment is made the long-term situation. In part this may explain the dis­ on the effect of this treatment on refraction. crepancy with other studies, all of which studied older Twenty-six other ocular findings were recorded at 6 infants, or children. We report, for the first time, that stra­ months. These were as follows, with numbers of infants in bismus is significantly associated with both the presence parentheses: nystagmus (4), mild (4), and stage of acute ROP, rising from 3.1 % in those without optic atrophy (3), double elevator palsy (2, in identical ROP to 29.2% with stage 3 disease. Kushnerlo also twins),26 disc coloboma (2), lens subluxation (1), cataract reported an increased incidence of strabismus in infants (1), ( I), microphthalmos (1) , eyelid haemangioma with regressed mild ROP compared with preterms who (1 ), simple (1), limbal dermoid (1 ), had not developed ROP. Infants with obvious abnormal­ hypoplasia (1), (1), pigmented (1) and ities of the central nervous system were excluded by homonymous hemianopia (1). Some infants had more Kushner, but there was no mention of neonatal neuro­ than one abnormality. logical evaluation, and he acknowledged that subtle 1.1 abnormalities may not have been detected. Schaffer et al. DISCUSSION found no major difference in the incidence of strabismus, There have been a number of studies reporting the oph­ and refractive errors in infants with regressed thalmic findings in infants and children born prematurely, mild (�stage 2) ROP and those who had not developed but as mentioned above, relatively few have compared retinopathy. Strabismus is a well-known associate of ophthalmic data from neonatal and subsequent periods. neurological abnormalities and several studies have We have recently studied prospectively a cohort of infants excluded abnormalities of the central nervous system in of birthweight � 1700 g (n = 572) commencing in the order to investigate the association between strabismus

Table VIII. Stage of acuteROP and axis of astigmatism for left eye in Table VI. Stage of acute ROP and mean spherical equivalent (in 387 infants refracted between 20 and 45 weeks corrected age. Numbers dioptres) for 387 infants refracted between 20 and 45 weeks corrected of infants for eachROP stage are given, with percentages in parentheses age Stage of ROP Right dioptres Left dioptres Stage of ROP (range) (range) Axis 0 2 3/4 Total o +1.85 +1.87 60°-120° 103 42 19 3 167 (-1.50 +7.50) to (-1.88 to +8.00) (66.5%) (55.3%) (40.4%) (23.1%) +1.91 +1.84 150°-30° 31 26 18 5 80 (-0.50 to +7.75) (-0.75 to +7.50) (20.0%) (34.2%) (38.3%) (38.450/() 2 +1.55 +1.76 Other 21 8 10 5 44 - (-6.25 to +4.00) ( 8 . 00 to +8.00) (l3.5c!',) (10.5%) (21.3%) (38.45%) 3/4 +0.45 +0.38 (-2.12 to +2.75) (-4.50 to +3.88) Total 155 76 47 13 291

With increasing ROP severity there is a trend towards myopia. from For infants who never developedROP the predominant axis is 60°_120°, stage 0 to 2. reaching significance only between stages 2 and 3 (right but with increasing ROP severity there was a significant trend away eye. p = 0.002; left eye. p = 0.003). Full ranges in parentheses. from this axis (p = 0.(02) and the proportion at other axes increased. Rap: SIX MONTH REVIEW 481 and ROP. However, these results must be interpreted with acuities in infants with regressed ROP (stage not given) caution for the method of exclusion is often relatively who had normal posterior poles - a result which persisted crude, as recognised by Kushner.lO Gibson et al.l7 in a after removal of infants with neurological abnormalities. prospective study of infants with a birthweight ::::;1500 g Luna et al.34 reported significantly delayed acuity who had regular cranial ultrasound examinations in the development in the first2 years of life in regressed stage 3 neonatal period noted an overall incidence of strabismus ROP, but in contrast to Birch and Spencer33 these workers of 10% (13/127), but 50% in those with cystic periventric­ concluded that this delay was due to neurological ular leucomalacia (7!14,p

31. Sebris SL, Dobson V, McDonald MA, Teller DY: Acuity 41. Linfield PB, Davis JG: A study of the change in mean sphere cards for visual assessment of infants and children in clinical refractive errors obtained from very low birthweight infants. settings. Clin Vis Sci1987, 2: 45-58. In: 1st International Congress, British College of Ophthal­ 32. Katsumi 0, Mehta MC, Matsui Y, Tetsuka H, Hirose T: mic Opticians (Optometrists), London 1984,pp. 197-200. Development of vision in retinopathy of prematurity. Arch 42. Seiberth V, Knorz MC, Trinkmann R: Refractive errors after OphthalmoI199l,109: 1394-8. cryotherapy in retinopathy of prematurity. Ophthalmologica 33. Birch EE, Spencer R: Visual outcome in infants with cicatri­ 1990,201: 5-8. cial retinopathy of prematurity. Invest Ophthalmol Vis Sci 43. Ben-Sira I, Nissenkorn I, Weinberger D, Shohat M, Kremer 1991,32: 410-15. I, Krickler R, Reisner SH: Long-term results of cryotherapy 34. Luna B, Dobson V, Biglan AW: Development of grating for active stages of retinopathy of prematurity. Ophthal­ acuity in infants with regressed stage 3 retinopathy of pre­ mology 1986,93: 1423-8. maturity. InvestOphthalmol Vis Sci 1990,31: 2082-7. 44. Cryotherapy for Retinopathy of Prematurity Cooperative 35. Sebris SL, Dobson V, Hartmann EE: Assessment and pre­ Group. Multicentre trial of cryotherapy for retinopathy of diction of visual acuity in 3- to 4-year-old children born prematurity. One year outcome - structure and function. prior to term. HumNew·ohiol 1984, 3: 87-92. ArchOphthalmol 1990; 108: 1408-16. 36. Fledelius HC: Ophthalmic changes from 10 to 18 years. A 45. Fielder AR, Moseley MJ, Ng YK: The immature visual longitudinal study of sequels to low birthweight. II. Visual system and premature birth. Br Med Bull 1988, 44: acuity. ActaOphthalmol 198l, 59: 64-70. 1093-118. 37. Fletcher MC, Brandon S: Myopia of prematurity. Am] Oph­ thalmol1955,40: 474-81. 46. Fledelius HC: Ophthalmic changes from 10 to 18 years. A 38. Graham MV, Gray OP: Refraction of premature babies' longitudinal study of sequels to low birthweight. III. Ultra­ eyes. BrMed ] 1963, i: 1452-4. sound oculometry and keratometry of the anterior segment. 39. Shapiro A, Yanko L, Nawratzki I, Merin S: Refractive power ActaOphthalmol 1982, 60: 393-402. of premature children at infancy and early childhood. Am] 47. Fledelius HC: Inhibited growth and development as per­ Ophthalmol 1980, 90: 234-8. manent features of low birth weight. Acta Paediatr Scand 40. Dobson V, Fulton AB, Manning K, Salem D, Petersen RA: 1982,71: 645-50. Cycloplegic refractions of premature infants. Am] Ophthal­ 48. Reynolds JD: Anisometropic amblyopia in severe posterior mol1981, 91: 490-5. retinopathy of prematurity. BinocVis Quart 1990, 5: 153-8.