The Importance of Timely Ophthalmologic Examination in Preterm Infants at Risk of Retinopathy Occurrence

Znaþaj pravovremenog oftalmološkog pregleda kod nedonošþadi sa prisutnim rizikom od pojave retinopatije Abstract Background/Aim. Retinopathy of prematurity (ROP) is a multifactorial disease in premature infants. The aim of this study was to determine the incidence of ROP in children treated at the Center of Neonatology, Pediatric Clinic, Clinical Center in Kragujevac, Serbia. Methods. The study covered all children with birth weight below 2,000 g and/or gesta-tional age below the 37th week, who from June 2006 to De-cember 2009 underwent ophthalmological examination for ROP. The results of fundoscopy were classified in accordance with the International Classification of ROP. The treatment of infants and those with ROP was conducted in accordance with the early treatment of ROP study recommendations. We analyzed gestational age, birth weight and postconceptional age in two groups: healthy infants and those with severe form of ROP. Statistical analysis was performed using the SPSS 16. Results. A total of 478 children met the criteria of screening for ROP. Severe stage of ROP, which required laser treatment, had 102 (21.3%) children. Out of the infants with severe ROP 14 (13.7%) of the infants with APD had aggressive posterior disease, while two (0.4%) remained blind. The differences in the mean values of gestational age between the healthy and the children with severe form of the disease were statistically significant (p < 0.0005). The mean value of gestational age for the healthy children was 33.33 r 2.28 weeks and for the seek infants 30.66 r 2.79 weeks. The mean value of the weight in healthy children was 1.981 r 407 g, and in sick children 1.535 r 434 g which was statistically significant (p < 0.0005). Multivariate binary logistic regression showed that the occurrence of the disease depends on body weight and gestational age. Conclusion. The incidence of severe forms of ROP was 21.3%. Aggressive form of ROP was present in 13.7% of the children. The cutoff value for body weight was 1.740 g, and for gestationl age 32.5 weeks.


Introduction
Retinopathy of prematurity (ROP) is a proliferative disease of immature retina that may cause severe sight damage or blindness [1][2][3] .Incidence of ROP and blindness in newborns depends on the level of care and methods used in neonatal intensive care units (NICU), and also more recently on economic development of a country 4 .
Two epidemics of ROP were recorded in developed countries in the past.The first one was in 1940s and 1950s, where the main causative factor was supplemental use of oxygen [4][5] .Second epidemic occurred in 1970s, due to increased survival rate of extremely prematurely born babies [4][5][6] .Recently, a "third epidemic" appeared in the developing countries 7 .Possible causes of the "third epidemic" were: increase in rate of preterm birth, increased survival rate of preterm babies due to development of NICUs; wider screening criteria.The babies are being exposed to less controlled risk factors than in developed countries.Like in the first epidemic, the babies affected in the third one were within wider range of birth weights (BW) and gestational ages (GA) 4,[8][9][10][11] .
Interestingly, both countries with high and low infant mortality rate (IMR) (> 60/1000 live births and < 9/1000 live births, respectively) do not record blindness caused by ROP, or it is very rare.In the countries with IMR between 9 and 60 per 1,000 live births, ROP emerges as an important cause of blindness 4 .In Serbia, the infant mortality rate is 14/1000 live births (national statistics for 2008), but there are no accurate data at the national level that would identify ROP as possible cause of blindness 12 .In order to prevent blindness due to ROP in countries like Serbia, ophthalmological screening of all newborns with risk of premature retinopathy should be conducted.However, screening criteria are different and subject to change in various countries 12 , and there exists the need at the national level to carefully review, revise and adapt these various screening guidelines to the context of the ROP situation in Serbia.
During the revision of international classification of ROP in 2005, the aggressive posterior disease (APD) was included in it 13 .The APD is more commonly seen in infants with BW less than 1,200 g, or with gestational age below 24 weeks.However, in the countries like Serbia, it is also present in infants with bigger BW and older than 24 weeks of GA.This disease is very aggressive, progresses rapidly and may cause retinal detachment and blindness.It requires rapid diagnostics and immediate treatment [12][13][14][15][16][17] .
The aim of this study was to estimate the incidence of ROP and to verify the applicability of national guidelines for screening in our country undergoing socio-economic transition.

Methods
The Neonatology Center of Pediatric Clinic in Kragujevac, Serbia, is an interregional center admitting high-risk infants from an area of 10 municipalities in Serbia.All infants admitted from June 2006 to December 2009 to the facility, with BW under 2000 g and/or GA under 37 weeks were submitted to screening 12 .Screening was conducted outside these criterions in cases with high risk of ROP development, such as respiratory distress.
The first ophthalmologic examination was done between the 4th and the 5th postnatal week.Fundoscopy was performed after dilatation of the pupils.We used 2.5% phenylephrine or 0.5% cyclopentolate drops three times per hour before examination, in order to achieve maximal dilatation of the pupils.Binocular indirect ophthalmoscope with condensing 20 D lens and an indenter was used for the examination.The results of fundoscopy were classified in accordance with the International Classification of Retinopathy of Prematurity (ICROP) 13 .This classification includes three parameters: localization of changes at retina, dilatation of pathologically changed blood vessels and stage of pathologically changed vascularisation.
After the first ophthalmological examination all newborns with normal findings were excluded from further follow-up.We continued to monitor newborns with undeveloped retinal blood vessels, but without signs of ROP, every 7 to 14 days.The treatment of newborns with ROP was conducted in accordance with Early Treatment of ROP (ETROP) study recommendations, which stressed necessity of discovering "prethreshold disease" and early treatment 13 .According to this study, ROP that needs treatment is type 1 ROP: zone I, any ROP stage with plus disease; zone I, stage 3 of ROP without plus disease; and zone II, stage 2 or 3 of ROP with plus disease 13 .Type 2 ROP requires follow-up visit: zone 1, stage 1-2 without plus, zone 2, stage 3 without plus.
During ophthalmological examination we searched for significant dilatation and tortuosity of both vessel types in the first zone, in all four quadrants.The APD appears much earlier than classical ROP, already after three weeks of postnatal age.The APD does not follow the classical pattern of stage 1 through stage 5.
We analyzed the post conceptual age (PCA) in all the groups.PCA is defined as actual postnatal age in weeks (days) minus difference of 40 weeks of gestation and gestation at birth.PCA refers to the time of screening, i.e. first ophthalmologic examination.All the children with AP ROP, the first exam marked the time of the first laser therapy.
Statistical analysis was performed using the Statistical Package for Social Sciences software, SPSS 16.The data are reported as means SD.The statistical significance of differences between the results was tested using a two-sided independent Student's t-test.The results were considered significantly different when p 0.05.The independency between categorical variables was tested by the 2 test.The optimal data cut-off was determined by the receiver operating characteristic (ROC) curves.Such cut-off was used for computing sensitivity and specificity.Risk of retinopathy was tested by the multiple binary logistic regression.

Results
The study population consisted of 1223 infants admitted to the Neonatology Center.From this primary population, 478 infants had risk factors for retinopathy.The study group consisted of 273 (57.1%) male and 205 (42.9%) female infants.After the first ophthalmologic examination, 131 (27.4%) infants had normally developed retinal blood vessels The other 347 (72.6%) infants had primarily undeveloped retinal blood vessels.Within this group that was subjected to further ophthalmological monitoring, 170 (35.6%) infants were found with insufficiently developed retinal blood vessels.They were added to the healthy group of 131 infants, and the total number of newborns with normally developed retinal blood vessels was 301 (62.9%).Seventy-five (15.7%) newborns developed stage 1 or 2 of ROP with spontaneous regression.They were added to the healthy infants, while 102 (21.3%) newborns developed severe stage of ROP, and underwent laser therapy (Figure 1).
The average values of GA, BW and PCA of 376 healthy infants and 102 infants with severe stage of ROP are shown in Table 1.
The differences in the mean values of gestational age between healthy and children with severe form of the disease were statistically significant (p < 0.010).The mean value of gestational age for the healthy children was 33.33 ± 2.28 weeks and for the seek infants 30.66 ± 2.79 weeks.Gestational age may be a marker (indicator) of the disease (retinopathy).The cut-off value was 32.5 weeks (Table 2).The sensitivity was 69.3% and the specificity 73.1%.The positive predictive value was 40.9 %, while the negative predictive value was 89.9 % (Figure 2).
The differences in the mean values of the weight between healthy and sick children were statistically significant (p < 0.0005).The mean value of the weight in the healthy children was 1,981 ± 407 g and in the sick children 1,535 ± 434 g.Birth weight in infants may be a marker of the disease (retinopathy) .The cut-off value was 1,740 g (Table 2).The sensitivity was 72.1%, and the specificity was 72.3%.The positive predictive value was 41%, and negative predictive value was 90.6% (Figure 3).Multivariate binary logistic regression showed that the occurrence of the disease depends on body weight (p < 0.0005) and gestational age (p = 0.010).The odds ratio  for weight was 0.998 (0.997-0.999).One gram of additional weight reduces the risk of developing the disease by 0.2%.The odds ratio for gestational age was 0.840 (0.736-0.959).One additional week of gestation reduces the risk of developing the disease by 16% (Table 3).
Out of 102 infants with severe stage of ROP 14 (13.7%)developedAPD.Female and male gender was equally present.The average GA of infants with APD was 28.7 (25 to 33) weeks.The average body weight was 1,307.4g (700 to 1,950) and when APD was diagnosed, PCA was 231.2 days (33 weeks).There were also two (0.4%) blind infants in the screened group, all of them with APD.Both infants had been laser treated immediately after the diagnosis of the disease, but the laser treatment was unsuccessful.

Discussion
ROP is a serious eye disorder that can lead to blindness 5 .Increased survival rate of premature infants requires screening and early treatment in order to control this disease 18 .The aim of screening is to detect all infants with ROP that require treatment.The CRYO-ROP study has defined the "threshold disease" of ROP as a stage when the treatment is necessary 19 .On the other hand, the ETROP study has shown that earlier treatment of highly risky "pre-threshold" stage of ROP leads to even better results 18 .
Many countries have national guidelines for screening in relation to BW and GA in accordance with studies on ROP incidence [20][21][22][23][24][25][26][27][28][29][30][31][32] .In developing countries (including Serbia) the incidence of ROP is higher, especially among premature infants with higher BW and GA [23][24] .In Serbia, screening is recommended even though the official protocol for screening does not exist.Recommended screening includes infants with GA below 37 weeks and/or BW of 2,000 g or less, as well as newborns at high risk outside this criteria, as discussed in this paper.
The incidence of severe ROP that required laser therapy in our study was 21.45% of all the screened infants.In the group of infants that had laser therapy, the minimal GA was 25 weeks and maximal 37 weeks.BW was in the range from 700 g to 3,200 g.Such a wide range of BW and GA is typical for developing countries, which fall under the "third epidemic" of ROP.
This shows that we should not just follow strictly recommended criteria since the aim of the screening is to discover all newborns at risk for severe ROP.The percentage of severe ROP does not follow closely the percentage of screening, which is high.The American Academy of Pediatrics and the American Academy of Ophthalmology recommend screening in infants with BW of less than 1,500 g or GA of 30 weeks or less, as defined by the attending neonatologist.It also included the selected infants with birth weight between 1,500 g and 2,000 g or GA greater than 30 weeks with an unstable clinical course, as well as those requiring cardiorespiratory support, as determined by their attending pediatrician or neonatologist.UK retinopathy of pre-

ROC Curve
Diagonal segments are produced by ties.

ROC Curve
Diagonal segments are produced by ties.
1 -Specificity   maturity guideline suggests the following screening criteria: all babies less than 32 weeks GA or less than 1,501 g BW should be screened for ROP and all babies less than 31 weeks GA or less than 1,250 g BW must be screened for ROP 12,24 .As an example of the need for careful adaptation of screening criteria to the local situation, if we had applied the UK screening criteria directly, we would have missed 31 infants with severe form of ROP, which required laser therapy.The majority of countries recommend screening in the 4th postnatal week, ie between 31 and 34 post conceptional week 19,23 .Because of early appearance of APD ROP, screening is recommended as early as 30-31 postconceptional week 31 .Screening before 30 weeks and laser therapy would be difficult to administer.

Conclusion
The incidence of severe form ROP in the studied population of our region is very high.Screening criteria applied in this work are much broader than in developed countries.We included all infants with BW under 2,000 g and/or GA under 37 weeks as well as infants outside these criteria in cases with high risk of ROP development.Applying the criteria from developed countries to our case series would have made us to omit a relatively large number of infants with severe form of ROP.In countries in transition, such is Serbia, the necessity for ROP screening is very high, as well as the need to create locally adapted screening criteria.

Fig. 3 -
Fig. 3 -ROC showing discriminatory value of birth weight (BW) for identifying newborn infants at risk for retinopathy of prematurity (ROP)