The effects of thyroid function on retinopathy of prematurity

Abstract

OBJECTIVES: To assess whether TSH and fT4 have a role in the angiogenesis of vaso-obliteration and neovascularization which are the basic pathophysiology of ROP.

METHODS: In this retrospective case-control study, the control group (n = 56) included preterm newborns with risk for ROP while the laser group (n = 63) was recruited from cases who developed severe neovascularization and needed laser photocoagulation therapy. Considering the first (vaso-obliteration) and second (neovascularization) phases of the disease, in this study we researched the distribution of thyroid function tests between groups.

RESULTS: With regard to the first phase of the disease, TSH and fT4 showed no significant differences between the control and laser groups accordingly (P > 0.05). Likewise, in the second phase of ROP, there was no significant difference between the control and laser groups with respect to TSH and fT4 levels (P > 0.05).

CONCLUSION: We found that between the study groups, the levels of thyroid function tests did not have any significant differences, either in the first or the second phases of ROP which are the principal pathophysiology of the disease. Therefore, it was concluded that thyroid hormone values were not informative markers in the course of the disease in preterm babies at risk of developing ROP.

Keywords

Retinopathy of prematurity thyroxine thyroid stimulating hormone vascular endothelial growth factor

Abbreviations

ROP

Retinopathy of prematurity

VEGF

Vascular endothelial growth factor

TSH

Thyroid-stimulating hormone (also known as thyrotropin)

fT4

Free thyroxine

RDS

Respiratory distress syndrome

BPD

Bronchopulmonary dysplasia

PROM

Premature rupture of membranes

IVH

Intraventricular hemorrhage

PDA

Patent ductus arteriosus

NEC

Necrotizing enterocolitis

1 Introduction

Retinopathy of Prematurity (ROP) is a developmental retinal vessel disease and one of the leading causes of childhood blindness all over the world [1]. Currently the most recognized hypothesis in the pathogenesis of the disease is the biphasic theory, which is first introduced by Aston and Smith [2, 3].

According to this theory, ROP develops in two phases. The first phase (vaso-obliteration) is marked with a decrease in the pro-angiogenic factors such as vascular endothelial growth factor (VEGF) and insulin-like growth factor-1 which have an effect on angiogenesis, while the second phase (neovascularization) is marked with increased production of the related mediators. The principal mechanism in the pathogenesis of ROP is excessive angiogenesis occurring in the second phase of the disease [3 –5].

In the literature, it was shown that both thyroid-stimulating hormone (also known as thyrotropin-TSH) and free thyroxine (fT4) have a pro-angiogenic effect on the vascularization through several pathways. These studies reported that thyroid hormones boost angiogenesis via an increment in VEGF, basic fibroblast growth factor, bradykinin, angiotensin-II, aminopeptidase and endothelial cell motility [6 –10].

According to the literature, the number of clinical studies done about the effect of thyroid hormones on angiogenesis is extremely low. Here we investigated whether thyroid hormones have an effect on the first phase, called vaso-obliteration, which in turn triggers the second phase, called neovascularization, which has principal role in the pathogenesis of ROP. To enable the aim of the study, we compared the subjects in early stage of ROP without need of laser therapy to the subjects in late stage of ROP with the need of laser therapy. Since the main difference between the two groups in the pathogenesis of ROP is the level of angiogenesis, detecting the difference among the effects of TSH, fT4 and pro-angiogenic factors in this respect is thought to be important. As known, the most significant difference between the newborns with need of laser therapy and without need of laser therapy was development of pathologic angiogenesis which is underlying mechanism for ROP [5]. Thus, it is important and valuable to search the difference between the groups in respect to thyroid hormone levels to understand the effect of thyroid hormones in the pathogenesis of ROP.

Since increased thyroid hormones in the second phase of the disease might contribute neovascularization by increasing VEGF, basic fibroblast growth factor, bradykinin, angiotensin-II, aminopeptidase, endothelial cell motility levels, detecting any significant difference in TSH-fT4 levels between groups in the second phase would be beneficial in the follow-up and ophthalmic screening of newborns at risk of ROP and may assist in deciding the optimal time for laser therapy [6, 9].

2 Methods

In this retrospective case-control study, control group (n = 56) included preterm newborns at risk for ROP while the other laser group (n = 63) were recruited from cases who developed severe neovascularization and needed laser photocoagulation therapy, and the study was approved by the Ethical Board of Erciyes University Faculty of Medicine.

Patients who were admitted to our hospital from August 2012 to January 2014. In our institution, newborns with a birth weight of less than 2000 grams and gestational age of less than 31 weeks are considered as newborns at risk for ROP. According to this definition, the control group was recruited.

Several studies have previously shown that mediators such as VEGF and IGF-1 are important in development of ROP and VEGF and IGF-1 levels were significantly increased among subjects in need of laser therapy compared to subjects without need of laser. Thus, we did not search the difference of the groups in respect to VEGF and IGF-1 [5, 11].

We included infants with birth weight <2000 g and gestational age <31weeks. This is the population that is considered at risk for ROP and receives routine ophthalmologic examination [12]. The laser group included the subjects with high-level neovascularization and as a result, the subjects in need of laser photocoagulation. On the other hand, the control group included subjects with low-level neovascularization who did not require laser photocoagulation. This type of enrollment made it possible for the assumption that the possible effect of thyroid function test values on neovascularization might be better differentiated between these two groups.

Grading ROP cases was done according to the International Classification of Retinopathy of Prematurity system: Stage 0: locating no sign of ROP located in the retina; Stage I: sighting of the thin whitish demarcation line which separates the avascular and vascular retinal areas from each other; Stage 2: sighting of the ridge characterized by the demarcation line gaining width and volume; Stage 3: defined as irregularities in the ridge and sighting within it the formation of new vessels. Plus Disease was defined as tortuosity and dilatation in the posterior retinal vessels [13]. In our study, the need of laser photocoagulation was determined in accordance with ET-ROP criteria [11, 14]. The laser group included 63 newborns who needed laser photocoagulation according to Early Treatment of ROP (ET-ROP) criteria.

The postpartum first week TSH and fT4 values of both groups were recorded and these values were used as values reflecting the first phase of ROP in these groups. Postmenstrual age was taken as a reference point in detection of the second phase in the study groups. None of TSH-fT4 values earlier than the 31st postmenstrual week were used in the analysis of any phase for ROP.

Postmenstrual 32nd week and earlier, with gestational age earlier than 27 weeks at the postmenstrual 33rd week and earlier with gestational age between 27–29 weeks, and at the postmenstrual 34th week and earlier with gestational age later than 30 weeks were recorded values of the TSH-fT4. These values were used as references reflecting the second phase of the disease.

In this study, TSH and fT4 levels were both evaluated together and measured on the same day. Thyroid hormones levels were measured using a Bayer Advia Centaur machine. Analyses of TSH and fT4 from undiluted serum samples of newborn were performed on an automated immunoassay system (Advia® Centaur™, Bayer Health Care Diagnostika, Vienna, Austria) using a direct chemiluminiscence detection system according to the manufacturer’s instructions [15, 16].

TSH was analyzed in a two-site solid-phase format, whereas analysis of fT4 was performed in a competitive assay. The intraassay coefficients of variation for TSH, fT3, and for fT4 were <8.7%, <3.5%, and <8.1% respectively. The reference values for adult patients using this immunoassay system have recently been published in BMC Endocrine Disorders and Thyroid [17, 18].

There are studies in the literature indicating that thyroid levels can be lower for 3 days following exchange transfusion [19]. In our clinical practice, we replace the need of blood transfusion with erythrocyte suspension instead of the whole blood. In case of exchange transfusion, blood sample for thyroid hormone was always drawn 3 days later after exchange transfusion in accordance with advice of pediatric endocrinology department. Thus, our results related to thyroid hormones were free of exchange transfusion. Since this point is very vital for the assessment of the results, this topic was mentioned in a different paragraph in the method section.

Cases to whom surfactant was given within 24 hours postpartum were recorded as having respiratory distress syndrome (RDS). Birth weights, gestational age in weeks, mechanical ventilation duration in hours and oxygen intake duration in days were recorded for all cases accordingly. Cases without TSH and fT4 values for the first and second phases of ROP were excluded from the study. An international scoring system was used in scoring bronchopulmonary dysplasia (BPD) [20].

Patients suspected of having BPD were included in the BPD records. Premature rupture of membranes (PROM) was considered positive when rupture of membranes occurred more than 18 hours before delivery. Patients with hemorrhage grade of at least 2 grades as scored by the Volpe scoring [21] method for intraventricular hemorrhage (IVH), were considered positive and thus recorded. Cases with a ratio of patent ductus arteriosus (PDA) over right atrium aortic root greater than 1.4 were considered positive. Cases with necrotizing enterocolitis (NEC) and sepsis were recorded together. The Bell classification was used for NEC [22, 23]. Patients who received points above 5 by Töllner criteria were recorded positive [24].

One hour before their examination, the infants’ pupils were enlarged with 1% tropicamide and 2.5% phenylephrine, each given 3 times at 15-minute intervals. Topical anesthesia was instituted with 0.5% proparacaine hydrochloride drop before the examination. With a binocular indirect ophthalmoscope following the attachment of its speculum, and by means of 20 and 28-diopter lenses, first the anterior segment and then the fundus were examined.

The data were analyzed by using the SPSS 16.0 (SPSS Inc. Chicago, Illinois) statistical package program. Distribution of the data was controlled via Shapiro-Wilk Normality test. Between groups, normally distributed variables were compared by using independent sample t test, and variables without normal distribution were compared by using the Mann-Whitney U test. The chi-square test was utilized to analyze rational data. A P value of <0.05 was accepted as statistically significant.

Exclusion criteria: subjects with congenital thyroid disease, subjects taking therapy for hypothyroidism, and newborns without clear documentations for thyroid hormones were excluded from the study.

3 Results

Fifty-six cases for the control group and sixty-three cases for the laser photocoagulation group were included in the study. The frequency of all diseases and interventions which could affect the course of ROP and the demographic properties of the groups are given in Tables 1 and 2 respectively. As mentioned in the literature [25], BPD and RDS are classified as the most important neonatal diseases to have an effect on the formation of ROP thus, in this respect, the study groups were examined closely (Table 2). The frequency of BPD was found to be 9 (16%) in the control group and 37 (58.7%) in the laser photocoagulation group (p: 0.001).

In case of RDS, its frequency was found to be 37 (66.0%) in the control group and 51 (80.9%) in the laser photocoagulation group (p: 0.04) (Table 2). IVH frequency was 30 (53.3%) in the control group while it was 16 (25.3%) in the laser photocoagulation group (p: 0.02) (Table 2). There was no significant difference found between the groups with respect to PROM, PDA, sepsis-NEC and preeclampsia (P > 0.05 and Table 2).

Gestational age, birth weight and gender, which are among the most important factors affecting ROP, were evaluated accordingly [26 –29]. The gestational age of the control group (28.0±1.9 weeks) did not statistically differ from that of the laser photocoagulation group (27.0±1.2 weeks) (p: 0.35) (Table 1).

In addition there was not a significant difference between the control group and the laser photocoagulation group with respect to birth weight (1070±301 gr) and (980±185 gr) respectively (p: 0.08) (Table 1).

One of the most important external interventions affecting the formation of ROP is duration of mechanical ventilation [5]. The duration of mechanical ventilation for the control group was (192.0±481.5 hours) and (888.0±496.8 hours) for the laser photocoagulation groups the difference was statistically significant (p: 0.01) (Table 1).

We did not find any statistically significant difference between the control group and the laser group (27.0±15.9 days) and (35.5±19.3 days) respectively, with respect to duration of oxygen intake (p: 0.34) (Table 1).

It was found that the TSH values of the control group in the first phase of ROP were not statistically significantly different when compared to those of the laser photocoagulation group (3.25±5.1μIU/mL) vs. (3.25±4.9μIU/mL) respectively (p: 0.72) (Table 3).

Similarly, the fT4 values were not statistically significant values between the control and laser groups (1.23±0.4 ng/dL) vs. (1.33±0.1 ng/dL) respectively (p: 0.91) (Table 3). Considering the TSH values of the groups in the second phase of ROP, the control group did not have statistically significant values compared to the laser photocoagulation group (3.39±1.4μIU/mL) vs. (3.41±1.6μIU/mL) respectively (p: 0.55) (Table 3). Similarly, the fT4 values in the second phase were not statistically significant in the control and laser groups (1.27±0.4 ng/dL) vs. (1.32±0.3 ng/dL) (p: 0.90) (Table 3).

Logistic regression analysis was used to evaluate whether TSH and fT4 levels were protective or risk factors for laser indication, and it was found that no statistical significance was present (Table 4).

4 Discussion

ROP is a two-phase disease, beginning with delayed retinal vascular growth after premature birth (Phase I). Phase II follows when phase I-induced hypoxia releases factors to stimulate new blood vessel growth. Both oxygen-regulated and non-oxygen-regulated factors contribute to normal vascular development and retinal neovascularization [5].

Many studies related to the pathophysiology of ROP, show that pro-angiogenic (such as; VEGF, insulin-like growth factor-I, basic fibroblast growth factor, bradykinin, angiotensin-II, aminopeptidase, endothelial cell motility) values are low in the first phase of the disease and higher in the second phase. Vascular endothelial growth factor is an important oxygen-regulated factor [5 , 30]. The increase of pro-angiogenic factors in the second phase of ROP as described above predicted the need of photocoagulation therapy in patients with ROP [5, 11].

According to the pathophysiology of ROP, it is known that a decrease in VEGF in the first phase and increase in the second phase initiate the pathologic angiogenesis responsible for the disease. In other words, according to the nature of the disease an excessively increased VEGF level in the second phase compared to the first phase induces the actual underlying pathology of ROP which in return results in patients progressing to the stage of laser surgery [5 , 31].

Studies related to the effect of thyroid hormones on angiogenesis, show that thyroid hormones boost angiogenesis by increasing pro-angiogenic factor levels in particular. Therefore, taking an assumption of a possible relation between thyroid hormone levels and ROP is quite logical. In this aspect, the contribution of TSH and fT4 levels on the increment of pro-angiogenetic factors in neovasculogenesis phase of ROP may be considered effective [6 –10].

We assumed that the best way to show the relation between thyroid hormones and ROP was to compare the newborn group in need of laser photocoagulation and the newborn group who are not in need. Therefore, we tested our hypothesis between these groups. Our results indicated that TSH and fT4 levels of newborns in need of laser photocoagulation didn’t differ statistically compared to the newborns without the need for laser. Thus further and more detailed studies are needed for better understanding on this issue. (Tables 3 and 4 and Figs. 1–4).

Being small for gestational age and having low birth weight are some of the major risk factors for ROP. At birth these two factors usually have a direct correlation with the retinal vasculature and immaturity of neural development [29].

No significant difference was found between the groups in terms of gestational age and birth weight and in this respect patients were distributed homogenously (Table 1). For obtaining optimal results, homogenous distribution of the factors effecting development of ROP was an important aspect of our study.

Oxygen is another important factor for ROP and this has been shown in a large-scale study [32, 33]. In the year 2010, one study showed that low oxygen saturation (70–96%) in the first few postnatal weeks and high oxygen saturation (94–99%) at postmenstrual 32 weeks and after reduced the risk of progression to severe ROP [26].

As a result of the study, it was concluded that reduction of oxygen application in the first phase and increment of oxygen application in the second phase could prevent development of ROP by inhibiting pro-angiogenic factors. The design of the study considering the different phases of the disease was an important factor. We also considered the different phases of ROP while searching the possible effects of thyroid hormones on ROP and on this aspect, our study resembles the other study.

No statistically significant difference was found between the control group and patient group with respect to oxygen use (Table 1). Considering the fact that one of the important factors on development of ROP is oxygen use, we minimalized the effect of oxygen use on the study results by reducing its usage, and this enabled better evaluation of our data.

In other studies, it was shown that development of ROP was increased due to impaired auto regulation of choroidal blood flow among newborns in need of mechanical ventilation [5].

In our study, the group with laser had longer mechanical ventilation time as expected (Table 1). Infants who suffered from severe diseases (BPD, PROM, sepsis-NEC, RDS, IVH, PDA, preeclampsia) are usually effected primarily by retinopathy of prematurity [25]. Many researchers have reported that an increase in oxygen intake in RDS patients results in increased risk of ROP development. As a result, growth factors inducing vasculogenesis in the first phase are suppressed and trigger the second phase which is the primary phase, responsible for the development of ROP [34, 35]. Several studies in many cases proved the effect of IVH, RDS, and BPD on ROP. Despite the fact that literature there are some studies in the literature showing the importance of these diseases in the etiology of ROP, except for BDP, there is still not a full consensus on these issues [25, 36].

In our study, BPD and RDS frequencies were significantly higher in the group with laser while IVH frequency was significantly higher in the control group without laser (Table 2). This was an expected result since BPD and RDS frequencies increase naturally among newborns in need of laser photocoagulation. Both the groups had statistically similar values in respect to PROM, PDA, sepsis-NEC, preeclampsia frequencies (Table 2). In fact, these data were related to the data of diseases whose effects on the pathogenesis of ROP have not been shown and are still in debate. Therefore, it was expected not to find statistically significant relation between ROP and these diseases.

As a result, TSH-fT4 levels may reflect levels of VEGF and other pro-angiogenic factors in premature newborns under risk for ROP. Our study was designed for the purpose of showing the increasing effect of thyroid hormones on angiogenesis, shown in other previous studies, among newborns who were under the risk of developing ROP [6 –10]. We believe that confirmation of such a relation may add a substantial contribution in follow-up of patients under the risk for ROP.

We did not find any statistical difference between the groups in respect to TSH-fT4 levels effecting the pro-angiogenic factors which were thought to reflect the first and second phase of ROP (Tables 3 and 4). In conclusion, large scale studies are needed to evaluate the role of TSH-fT4 in follow-up of ROP and until the results of larger scale studies are obtained and at this point, we suggest that the use of thyroid hormones in follow-up of ROP may not be beneficial.

The limitations of our study; not multicenter and VEGF levels were not measured. But, there are a limited number of studies in the literature on this topic.

5 Conclusion

The only method used in the follow-up of newborns under risk for ROP is examination by indirect ophthalmoscope. However to both the physicians and the patients, this ophthalmologic examination method is time-consuming, quite difficult to perform and has also several potential side effects. Thus, we aimed to research whether thyroid hormone tests can be used in the follow-up of ROP among high-risk newborns and therefore designed our study accordingly.

In our study, we could not obtain significant results, which would be helpful to both patients and medical staff in the follow-up and ophthalmic screening of the disease, with respect to TSH-fT4 levels reflecting increased pro-angiogenic factor levels in the neovascularization phase. Therefore, until the results of meta-analyses on a larger scale multi-centered studies are carried out, we concluded that it would be more appropriate not to use thyroid hormones in follow-up of patients with ROP.

Conflicts of interest

Korkmaz L, None; Baştuğ O, None; Daar G, None; Korkut S, None; Özdemir A, None; Öztürk MA, None; Güneş T, None; Kurtoğlu S, None.

References

Gilbert

. Retinopathy of prematurity: A global perspective of the epidemics, population of babies at risk and implications for control. Early Hum Dev 2008;84:77–82.

Silva

, Moshfeghi

. Interventions in retinopathy of prematurity. Neoreviews 2012;13:e476–85.

Chen

, Joyal

, Hatton

, et al. Propranolol inhibition of β-adrenergic receptor does not suppress pathologic neovascularization in oxygen-induced retinopathy. Invest Ophthalmol Vis Sci 2012;53:2968–77.

Italiano

Jr , Richardson

, Patel-Hett

, et al. Angiogenesis is regulated by a novel mechanism: Pro- and antiangiogenic proteins are organized into separate platelet alpha granules and differentially released. Blood 2008;111:1227–33.

Sapieha

, Joyal

, Rivera

, et al. Retinopathy of prematurity: Understanding ischemic retinal vasculopathies at an extreme of life. J Clin Invest 2010;120:3022–32.

Carmona-Cortés

, Rodríguez-Gómez

, Wangensteen

, et al. Effect of thyroid hormone-nitric oxide interaction on tumor growth, angiogenesis, and aminopeptidase activity in mice. Tumour Biol 2014;35:5519–26.

Liu

, Zheng

, Shi

, Yuan

, Li

. Thyroid hormone induced angiogenesis through the integrin αvβ3/protein kinase D/histone deacetylase 5 signaling pathway. J Mol Endocrinol 2014;52:245–54.

Mousa

, Lin

, Tang

, Hercbergs

, Luidens

, Davis

. Modulation of angiogenesis by thyroid hormone and hormone analogues: Implications for cancer management. Angiogenesis 2014;17:463–9.

, Teng

, Jiang

. Relationship between preoperative serum TSH levels and expression of VEGF in papillary thyroid carcinoma. Asia Pac J Clin Oncol 2014;10:149–52.

10.

Silva

, Ocarino

, Vieira

, Nascimento

, Serakides

. Effects of hypo-and hyperthyroidism on proliferation, angiogenesis, apoptosis and expression of COX-2 in the corpus luteum of female rats. Reprod Domest Anim 2013;48:691–8.

11.

Kermorvant-Duchemin

, Sapieha

, Sirinyan

, et al. Understanding ischemic retinopathies: Emerging concepts from oxygen-induced retinopathy. Doc Ophthalmol 2010;120:51–60.

12.

Heller

, O’Shea

, Yao

, et al. Human milk intake and retinopathy of prematurity in extremely low birth weight infants. Pediatrics 2007;120:1–9.

13.

International Committee for the Classification of Retinopathy of Prematurity. The international classification of retinopathy of prematurity revisited. Arch Ophthalmol 2005;123:991–9.

14.

Phelps

, ET-ROP Cooperative Group. The early treatment for retinopathy of prematurity study: Better outcomes, changing strategy Pediatrics 2004;114:490–1.

15.

Kapelari

, Kirchlechner

, Högler

, Schweitzer

, Virgolini

, Moncayo

. Pediatric reference intervals for thyroid hormone levels from birth to adulthood: A retrospective study. BMC Endocr Disord 2008;8:15.

16.

Rawlins

, Roberts

. Performance characteristics of six third-generation assays for thyroid-stimulating hormone. Clin Chem 2004;50:2338–44.

17.

Moncayo

, Moncayo

, Virgolini

. Reference values for thyrotropin. Thyroid 2005;15:1204–5.

18.

Moncayo

, Dapunt

, Moncayo

. Diagnostic accuracy of basal TSH determinations based on the intravenous TRH stimulation test: An evaluation of tests and comparison with the literature. BMC Endocr Disord 2007;7:5.

19.

Mahyar

. Effects of exchange transfusion on neonatal thyroid function. Indian J Pediatr 2007;74:511–2.

20.

Bhandari

, Panitch

. Pulmonary outcomes in bronchopulmonary dysplasia. Semin Perinatol 2006;30:219–26.

21.

Bassan

, Benson

, Limperopoulos

, et al. Ultrasonographic features and severity scoring of periventricular hemorrhagic infarction in relation to risk factors and outcome. Pediatrics 2006;117:2111–8.

22.

Gregory

, Deforge

, Natale

, Phillips

, Van Marter

. Necrotizing enterocolitis in the premature infant: Neonatal nursing assessment, disease pathogenesis, and clinical presentation. Adv Neonatal Care 2011;11:155–64; quiz 165-6.

23.

Lin

, Shi

, Luo

, Bao

, Ma

. Value of radiographic assessment scale in necrotizing enterocolitis. Zhongguo Dang Dai Er Ke Za Zhi 2012;14:97–100.

24.

Altunhan

, Annagür

, Örs

, Mehmetoğlu

. Procalcitonin measurement at 24 hours of age may be helpful in the prompt diagnosis of early-onset neonatal sepsis. Int J Infect Dis 2011;15:854–8.

25.

Leviton

, Dammann

, Engelke

, et al. The clustering of disorders in infants born before the 28th week of gestation. Acta Paediatr 2010;99:1795–800.

26.

Chen

, Guo

, Smith

, Dammann

. High or low oxygen saturation and severe retinopathy of prematurity. Pediatrics 2010;125:1483–92.

27.

Castillo

, Deulofeut

, Critz

, Sola

. Prevention of retinopathy of prematurity in preterm infants through changes in clinical practice and SpO2 technology. Acta Paediatr 2011;100:188–92.

28.

Akman

, Demirel

, Yenice

, Ilerisoy

, Kazokoglu

, Ozek

. Screening criteria for retinopathy of prematurity in developing countries. Eur J Ophthalmol 2010;20:931–7.

29.

Darlow

, Hutchinson

, Henderson-Smart

, et al. Prenatal risk factors for severe retinopathy of prematurity among very preterm infants of the Australian and New Zealand Neonatal Network. Pediatrics 2005;115:990–6.

30.

Sapieha

, Sirinyan

, Hamel

, et al. The succinate receptor GPR91 in neurons has a major role in retinal angiogenesis. Nat Med 2008;14(10):1067–76.

31.

Alon

, Hemol , Itin

, Pe’er

, Stone

, Keshet

. Vascular endothelial growth factor acts as a survival factor for newly formed retinal vessels and has implications for retinopathy of prematurity. Nat Med 1995;1:1024–8.

32.

Quinn

, Gilbert

, Darlow

, Zin

. Retinopathy of prematurity: An epidemic in the making. Chin Med J (Engl) 2010;123:2929–37.

33.

SUPPORT Study Group of the Eunice Kennedy Shriver NICHD Neonatal Research Network, Carlo

, Finer

, et al. Target ranges of oxygen saturation in extremely preterm infants. N Engl J Med 2010;362(21):1959–69.

34.

Chen

, Connor

, Aderman

, Willett

, Aspegren

, Smith

. Suppression of retinal neovascularization by erythropoietin siRNA in a mouse model of proliferative retinopathy. Invest Ophthalmol Vis Sci 2009;50:1329–35.

35.

Filippi

, Cavallaro

, Fiorini

, et al. Study protocol: Safety and efficacy of propranolol in newborns with Retinopathy of Prematurity (PROP-ROP): ISRCTN18523491. BMC Pediatr 2010;10:83.

36.

Farooqi

, Hagglof

, Sedin

, Serenius

. Impact at age 11 years of major neonatal morbidities in children born extremely preterm. Pediatrics 2011;127:1247–57.