Ocular outcomes and comorbidities in preterm infants enrolled for retinopathy of prematurity screening: A cohort study from western India.

BACKGROUND: Retinopathy of prematurity (ROP) is emerging as an important cause of childhood blindness in middle-income countries such as India. Although blindness can be prevented in most cases with timely screening and treatment, certain ocular comorbidities can lead to visual impairment. We evaluated and compared 1-year visual, refractive, and structural outcomes and comorbidities in three subsets of preterm infants enrolled for screening of ROP. SUBJECTS AND METHODS: Preterm children enrolled in the hospital's ROP screening program and with diagnosis of no ROP, mild ROP, or severe ROP were followed at 1 year of age to evaluate and compare visual, refractive, and structural outcomes as well as the presence of ocular comorbidities. Risk of poor outcome was calculated in children with mild and severe ROP reference population being children without ROP.
RESULTS: Eyes with severe ROP were at highest risk of poor visual (risk ratio [RR]: 3.5, P < 0.001), refractive (RR: 45, P < 0.001), and structural (RR: 11, P = 0.006) outcome as well as ocular comorbidities (RR 11, P < 0.001). Eyes with mild ROP were at higher risk of myopia (RR: 11, P = 0.06) and ocular comorbidities (RR: 4, P = 0.04). Sixteen (16%) of the eyes without ROP developed poor visual outcome.
CONCLUSION: Eyes with severe ROP are at highest risk of poor ocular outcomes and comorbidities and need a long-term follow-up. Eyes which do not develop ROP can have poor visual outcome and need to be assessed within the 1st year of life.

Introduction

Retinopathy of prematurity (ROP) is a potentially blinding retinal disease occurring principally in preterm low birth weight (BW) infants. Middle-income countries such as India are afflicted with the “third epidemic” of ROP.[1] Although prevention of blindness is possible with timely screening and treatment in most cases, certain ocular conditions can lead to moderate/severe visual impairment despite successful treatment for ROP. Prematurity can itself be associated with ocular comorbidities such as myopia, cataract, strabismus, anisometropia, nystagmus, and cerebral causes leading to visual impairment. Poor compliance for long-term follow-up in this group could be a factor responsible for delayed diagnosis and poor visual outcome.[2] Laser treatment for ROP has been established as a risk factor for myopia.[3] There are studies from India and elsewhere reporting refractive/structural outcomes in infants with or without ROP.[345678] Comparison of the risk of poor visual, refractive, and structural outcomes and ocular comorbidities in three subsets of preterm infants, namely, those with: (1) no ROP, (2) mild ROP, and (3) severe ROP has not been reported from India. To the best of authors’ knowledge, there are very few studies comparing visual outcomes in preterm children with and without ROP.[91011] This could be due to difficulty in measuring visual acuity in infants owing to the need of specialized equipment and skilled human resource.

Rationale

Analyzing visual, refractive, and structural outcome for aforementioned subsets of preterm infants would possibly identify the risks of poor ocular outcomes in each of them. Comparison of these risks can help in customizing follow-up protocols separately for each category if required. This evidence-based protocol might obviate the need for frequent follow-up visits and improve compliance and hence visual outcomes in these children. This study was planned with the purpose of evaluating and comparing 1-year visual, refractive, and structural outcomes as well as other comorbidities and calculating the risk of poor outcomes in three-subset preterm infants, namely, those without ROP and with mild ROP and severe ROP.

Subjects and Methods

This cohort study was carried out in a tertiary eye care center in Pune, western India, between June 2013 and October 2014. Institutional Ethics Committee approval was obtained and Helsinki protocol was adhered to. Preterm infants undergoing ROP screening under study hospital's ROP program were enrolled historically and concurrently in the study. Gestational age (GA) and BW cutoff for ROP screening are 34 weeks and 2000 g, respectively. Digital wide-field imaging was performed by a trained technician and diagnosis was done by a ROP specialist. Modified International Classification of ROP (ICROP) was used for classifying the disease.[12] Disease was treated as per Early Treatment for ROP (ETROP) guidelines.[13]

Disease definitions

“Mild ROP” was defined as ROP which regressed spontaneously without treatment. “Severe ROP” was defined as ROP requiring laser treatment. Group of children who did not develop ROP (“no ROP” group) served as the reference population.

Sample size calculation

Assuming that 50% of laser-treated infants and 20% of untreated infants were likely to develop poor outcomes, with 5% error alpha (confidence interval [CI]: 95%), power of 80, and distribution of 1:1 in each group, total sample size calculated was 150 (300 eyes, 50 infants in each group). Considering 10% loss to follow-up, 326 eyes of 163 infants were enrolled. Of these, 150 infants (300 eyes) completed follow-up and could be analyzed for above-mentioned outcomes at 1 year of age.

Assessment of outcomes

All participants underwent visual, refractive assessment by a trained and experienced pediatric optometrist at 1 year of age. Examination for structural outcome and ocular comorbidities was done by an ophthalmologist.

Visual acuity assessment was performed using Teller Acuity Cards. Snellen equivalent of visual acuity of “< 6/48” was defined as “poor visual outcome.”[1415] Cycloplegic refraction was performed to assess refractive status. Myopia of “−2.00D or more” was defined as “poor refractive outcome.” Structural assessment was done using RetCam and indirect ophthalmoscopy. Any retinal sequelae such as retinal fibrous scar, tractional retinal detachment, disc/macular drag, or straightening of the major vascular arcades were defined as “poor structural outcome.” Similarly, assessment for ocular comorbidities was performed by an ophthalmologist using torch and handheld slit lamp. Pathologies such as nystagmus, strabismus, anisometropia, amblyopia, and cataract were defined as ocular comorbidities. All the aforementioned assessments are done as a routine protocol for examination of preterm children. This ensured the quality of data obtained historically.

Statistical analysis

Data were entered into an excel spreadsheet, and (StataCorp LLC, Texas, USA) was used for the analysis. Group of preterm children without ROP served as the reference population to calculate the risk ratios (RRs) of poor outcomes in other two groups.

Results

Visual outcome was poor in 16 (16%) of no ROP eyes, 23 (23%) of mild ROP eyes, and 56 (56%) of severe ROP eyes. RR for poor visual outcome was 3.5 (2.3–5.3, P < 0.001) in severe ROP cases. Poor visual outcome was attributed to either poor refractive/structural outcome or other ocular comorbidities. Causes of poor visual outcome in each group of infants are shown in Table 1. When there was more than one cause present, most treatable cause was considered as responsible for poor visual outcome in that infant.

Table 1

Causes of poor visual outcome in each subset of study population

Cause	No ROP (%)	Mild ROP (%)	Severe ROP (%)
Myopia	0	11 (47.8)	45 (80.3)
Strabismus/nystagmus	4 (25)	5 (21.7)	0
Retinal sequelae	0	0	11 (19.7)
Delayed visual maturation	10 (62.5)	4 (17.4)	0
Cortical causes	2 (12.5)	3 (13)	0
Total	16 (100)	23 (100)	56 (100)

ROP: Retinopathy of prematurity

Refractive outcome

Myopia developed in nearly half of eyes with severe ROP and a tenth of eyes with mild ROP. None of the eyes with “no ROP” developed myopia [Table 2].

Table 2

Refractive outcome in study population

Myopia	No ROP (%)	Mild ROP (%)	Severe ROP (%)
Yes	0	11 (11)	45 (45)
No	100 (100)	89 (89)	55 (55)
Total	100 (100)	100 (100)	100 (100)

ROP: Retinopathy of prematurity

RR for development of myopia was 45 (33–56, P < 0.001) in severe ROP eyes and 11 (4–17, P = 0.006) in “mild ROP” eyes.

Binary logistic regression was performed to calculate effect of other confounders such as BW, GA, and number of laser spots. This showed a very strong evidence that GA of <28 weeks (P = 0.002) and more than 2000 laser spots (P < 0.001) were associated with development of myopia.

Further stratification of severe ROP cases was done by GA (<28 weeks) and number of laser spots (>2000) to calculate the risk of myopia among severe ROP cases exposed to these factors. Table 3 shows distribution of myopia among children born with GA of less and more than 28 weeks.

Table 3

Refractive outcome among severe retinopathy of prematurity cases by gestational age

Myopia	GA		Total (%)
	<28 weeks (%)	>28 weeks (%)
Yes	26 (51)	22 (44.9)	48 (48)
No	25 (49)	27 (55.1)	52 (52)
Total	51 (100)	49 (100)	100 (100)

GA: Gestational age

RR for development of myopia among severe ROP cases with GA of <28 weeks was 1.1 (95% CI: 0.7–1.7, P = 0.5). Similarly, stratification of severe ROP cases by number of laser spots [Table 4] showed that RR for the development of myopia among severe ROP cases receiving >2000 laser spots was 1.4 (95% CI: 0.9–2.2, P = 0.15).

Table 4

Refractive outcome among severe retinopathy of prematurity cases by number of laser spots

Myopia	Laser spots		Total
	>2000 (%)	<2000 (%)
Yes	25 (50)	18 (36.7)	43 (43.4)
No	25 (50)	31 (63.4)	56 (56.6)
Total	50 (100)	49 (100)	99 (100)*

*Missing value - 1

Structural outcome

Assessment of structural outcome showed that 11% of “severe ROP” eyes had poor structural outcome in the form of retinal sequelae and RR for the same was 11 (4–17, P = 0.006) in this group. Eyes with diagnosis of aggressive posterior ROP (APROP) had a higher risk of retinal sequelae (37% vs. 9%, P = 0.05). None of the children from mild ROP group showed any evidence of retinal sequelae.

Ocular comorbidities

Ocular comorbidities were present in 21 (21%) of severe ROP eyes and 4 (4%) of mild ROP eyes. Risk was greatest in severe ROP eyes (RR: 21, 95% CI: 12–29, P < 0.001) and less in “mild ROP” eyes (RR: 4, 95% CI: 1–7, P = 0.04).

Discussion

Some children from each of the three subgroups recorded poor visual outcomes, although in different proportions. Under fifth (16%) of eyes which did not develop ROP still had poor visual outcome at 1 year of age. Delayed visual maturation was the most common cause of poor visual outcome in this group. The proportion of children with poor visual outcome was higher in other groups – nearly quarter of eyes from mild ROP group and over half of the eyes from severe ROP group developed it. The most common cause was myopia. Eyes with severe ROP were 3.5 times more likely to have poor visual outcome than the eyes without ROP. This could be due to high incidence of myopia in this group. These findings establish the need for visual acuity assessment during the 1st year of life in all preterm infants whether they develop ROP or not. Early detection of visual problem might enable early intervention in this group, thus achieving optimal visual outcomes. Possibility of poor visual outcomes irrespective of the development of ROP in preterm low BW children has been reported in few earlier studies as well.[1011]

Not only eyes with severe ROP but also the eyes with mild ROP were at higher risk of developing myopia at the end of the 1st year although the risk was much higher in eyes with severe ROP. This indicates that any form of ROP is an independent risk factor for myopia development in preterm children and underlines the need for close follow-up for this group. None of the eyes without ROP carried a documented risk of myopia. In a report of prevalence of myopia in the children enrolled in ETROP study, risk of myopia and high myopia at corrected age of 9 months was similar in children managed with early treatment and those managed conventionally.[16] However, both these groups included only treatable ROP cases. The definition of myopia (>−0.25D) and high myopia (>−5.0D) used in ETROP study was different than the present study. The incidence of myopia might likely have been underestimated in the present study owing to higher cutoff (>−2.0).

Other confounders such as lower GA and higher number of laser spots were strongly associated with risk of myopia. In lower GA (<28 weeks) group, there was very little evidence (P = 0.5) to suggest that severe ROP was an independent risk factor for myopia. However, there was suggestive evidence (P = 0.15) of role of higher number of laser spots (>2000) as an independent risk factor for the development of myopia. Risk of myopia attenuated in severe ROP cases after accounting for effects of GA and number of laser spots. Studies from the United States of America,[17] Taiwan,[4] and China[7] have reported association between refractive outcomes and ROP. In a study from Taiwan,[4] only laser-treated preterm children developed myopic refractive error and other groups, namely, mild ROP, no ROP, and full-term children did not develop significant myopia at all. In the present study, children with mild ROP were also at higher risk of myopia than those without ROP. This could be because of other factors such as lower GA which was shown to be a risk factor for myopia. In a study from China,[7] it was reported that preterm children with or without ROP were at higher risk of development of myopia and astigmatism compared to full-term children. This finding could be due to role of possible genetic factors in Asian, particularly Chinese population as is reported in some studies.[1819] None of the aforementioned studies reported visual and anatomical outcomes/other comorbidities in their study population.

Eyes with “severe ROP” were at 11 times the risk of developing retinal sequelae than eyes with “no ROP” or “mild ROP.” There is quite strong evidence to suggest that eyes with APROP have a higher risk of retinal sequelae compared to those with classical ROP. This could be due to particular aggressive nature and poor prognosis in this disease as is mentioned in a study from India[20] and vast area of the retina that needs to be lasered in these cases. The presence of retinal sequelae underlines the potential for suboptimal vision development and future risk of retinal detachment in this subgroup and hence the need for lifelong follow-up.

Although eyes with “severe ROP” were at highest risk (21 times) of developing ocular comorbidities, those with mild ROP were also at significant (4 times) risk compared to those without any ROP. This establishes the need for ocular screening during the 1st year of life even in children with mild ROP.

One of the limitations of this study is that the role of other risk factors for ROP (such as respiratory distress, oxygen supplementation, intraventricular hemorrhage, and poor weight gain) in the development of poor visual/refractive outcomes was not studied. These factors could be potential confounders making the findings of this study biased.

Conclusion

Risk of poor visual and refractive outcomes is high in children with mild ROP and highest in those with severe ROP compared to those without ROP. Risk of poor structural outcome is exclusive to severe ROP group. Ocular comorbidities are more common in children with any ROP. Preterm children who do not develop ROP should also be screened for visual outcome during the 1st year of life. Preterm children developing any stage of ROP irrespective of whether treatment was needed or not should be followed up closely to detect and treat any refractive error or ocular comorbidities. Children with severe ROP need lifelong follow-up.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.