Journal of Perinatology (2014) 34, 453–457 © 2014 Nature America, Inc. All rights reserved 0743-8346/14 www.nature.com/jp ORIGINAL ARTICLE A prospective study on hyperglycemia and retinopathy of prematurity L Mohsen1, M Abou-Alam1, M El-Dib2, M Labib1, M Elsada3 and H Aly2 OBJECTIVE: Retinopathy of prematurity (ROP) constitutes a significant morbidity in premature infants that can lead to blindness. Multiple retrospective studies have identified neonatal hyperglycemia as a risk for developing ROP. However, in the absence of any reported prospective study, it is not clear whether hyperglycemia is associated with ROP independent of the commonly associated comorbidities. The objective of this study was to investigate whether hyperglycemia in premature infants is independently associated with ROP. STUDY DESIGN: Premature infants (o1500 g or ⩽ 32 weeks gestational age) were enrolled in a prospective longitudinal cohort study. All demographic, clinical and laboratory data were collected. Bedside whole-blood glucose concentration was measured every 8 h daily for 7 days. For any glucose readingo50 or>150 mg dl − 1, serum sample was sent to the laboratory for confirmation. Hyperglycemia was defined as any blood glucose level ⩾ 150 mg dl − 1. ROP patients were compared with non-ROP patients in a bivariate analysis. Variables significantly associated with ROP were studied in a logistic regression model. RESULT: A total of 65 patients were enrolled with gestational age 31.1 ± 1.2 weeks and birth weight 1385 ± 226 g. Thirty-one patients (48%) were identified with hyperglycemia. On eye examination, 19 cases (29.2%) had ROP (13 with stage 1, 4 with stage 2 and 2 with stage 3). There were more cases of ROP in the hyperglycemia group compared with the euglycemia group (45% vs 15%, P = 0.007). Patients who developed ROP had significantly higher maximum and average glucose concentrations when compared with non-ROP patients. Multiple factors have been associated with ROP on bivariate analysis, including gestational age, exposure to oxygen, respiratory support and poor weight gain. However, in a logistic regression model including all significant variables, average blood glucose in the first week of life was the factor independently associated with ROP with an odds ratio of: 1.77 (95% confidence interval: 1.08 to 2.86), P = 0.024. CONCLUSION: In a cohort of premature infants, elevated average blood glucose concentrations in the first week of life is independently associated with the development of ROP. Journal of Perinatology (2014) 34, 453–457; doi:10.1038/jp.2014.49; published online 27 March 2014 INTRODUCTION hyperalimentation,15,16 blood transfusions11,17 and vitamin E 18 Hyperglycemia is a significant risk factor for morbidity and deficiency. Moreover, ROP has been associated with the 19 mortality in preterm infants. There are multiple physiological presence of patent ductus arteriosus, bronchopulmonary and biochemical mechanisms in preterm infants that can lead to dysplasia,19 sepsis,16,20 systemic candidiasis21 and intraventricular excess glucose production, insulin resistance or glucose intoler- hemorrhage.22 ance; the sequelae of these disturbances in glucose metabolism Several retrospective studies have recently suggested hyper- – are extensive.1,2 Hyperglycemia could be associated with osmotic glycemia as a possible risk factor for ROP.23 27 Hyperglycemia is diuresis and dehydration, which increase the risk of cerebral commonly associated with many conditions in very low birth bleeding and electrolyte imbalance. There are also concerns that it weight infant, including sepsis, candidiasis, intraventricular may be associated with increased mortality and other morbidities, hemorrhage and postnatal steroids, all of which are frequently such as decreased immunity, increased infection, poor wound encountered in infants who later develop ROP.23 Meanwhile, it is healing and loss of skeletal and cardiac muscles.3,4 quite plausible that hyperglycemia imposes biological changes to Retinopathy of prematurity (ROP) remains a leading cause the retina; adults with poorly controlled diabetes develop a of morbidity in very low birth weight infants.5 The provision distinct neoproliferative retinopathy that is particularly identified of supplemental oxygen, while a major risk factor, is not the in patients with elevated serum glucose over prolonged periods.28 only cause of the disease as demonstrated by a series of low There has been a consensus agreement for the need to birth weight infants who developed ROP without ever receiving prospectively test the association of hyperglycemia with – any oxygen.6 Other proposed risks for the development of ROP ROP.29 31 To the best of our knowledge, we are not aware of include poor postnatal growth,7,8 hypoxia,9 hypercarbia,10 any prospective study to address such an association that is hypocarbia,11 exposure to prolonged and aggressive mechanical critically important, especially with the growing interest in early ventilation,12 inotrope therapy,6 postnatal steroids,13,14 prolonged and aggressive parenteral nutrition. The aim of this prospective 1Department of Pediatrics, Faculty of Medicine, Cairo University, Cairo, Egypt; 2Department of Neonatology, The George Washington University and Children’s National Medical Center, Washington, DC, USA and 3Department of Ophthalmology, Faculty of Medicine, Cairo University, Cairo, Egypt. Correspondence: Dr H Aly, Newborn Services, The George Washington University Hospital, 900 23rd Street, NW, Suite G-2092, Room G-132, Washington, DC 20037, USA. E-mail: [email protected] Received 25 November 2013; revised 14 February 2014; accepted 14 February 2014; published online 27 March 2014 Hyperglycemia and ROP L Mohsen et al 454 cohort study was to determine whether elevated glucose RESULTS concentration is an independent risk factor for the development The study was conducted on 65 preterm neonates admitted to the of ROP in premature infants. NICU in Cairo University during a period of 6 months; out of the 65 enrolled subjects, 31 (48%) were hyperglycemic. The euglycemia and hyperglycemia groups did not differ in gestational age, birth METHODS weight or mode of delivery. Infants in the hyperglycemia group Patients were more likely to be females, had greater acuity of illness using We conducted a prospective cohort study on premature neonates the Clinical Risk Index for Babies Score, with medians (ranges) of admitted to the neonatal intensive care unit (NICU) at Cairo University 1 (0 to 6) vs 2 (0 to 7), P = 0.001, and received more frequent red Children’s Hospital, which is the largest referral tertiary care unit in the cell transfusions (Table 1). In the hyperglycemia group, there was country. Infants were included in the study if admitted within 24 h of life no significant difference between the mean bedside glucose vs − with gestational age (GA) ⩽ 32 weeks or birth weighto1500 g. Infants with serum laboratory testing; 267.6 ± 101.2 vs 250 ± 86.4 mg dl 1, major congenital anomalies were excluded from the study. The study was P = 0.46. approved by the ethics committee and was conducted in accordance with Forty-six infants (70.8%) had normal fundus examination, the University bylaws for human research. Parental consents were while 19 cases (29.2%) had ROP. Out of the 19 ROP cases, 13 obtained for all subjects. had stage-1, 4 infants had stage-2, and 2 infants had stage-3. Full maternal and perinatal history was collected for all the studied Compared with non-ROP infants (n = 46), cases with ROP (n = 19) neonates. Other NICU data such as the use of oxygen, ventilation and fi o phototherapy were documented. Data on date of start and day of full were signi cantly of younger gestational age (P 0.001), smaller establishment of enteral feeds, type of feeding and full caloric intake were birth weight (Po0.001), more likely to have respiratory distress recorded. (P = 0.001) and exposed to phototherapy (P = 0.007). Maternal and neonatal factors associated with ROP are presented in Table 2. Eye examination There were more cases of ROP in the hyperglycemia group Fundus examination was initially done at 4 to 6 weeks in compliance 32 compared with the euglycemia group (45% vs 15%, P = 0.007), with the recommendations of American Academy of Pediatrics. odds ratio 4.776 (95% confidence interval (CI): 1.46 to 15.60). Follow-up examinations were scheduled during the NICU stay and after fi Figure 1 presents average daily glucose concentrations in infants hospital discharge as based on retinal ndings. Examinations were fi performed by experienced pediatric ophthalmologists while masked with and without ROP during the rst week of life. Mean, to the glycemia condition of the infants. Retinal examination was maximum and minimum values for glucose concentrations in done using binocular indirect ophthalmoscope, lid speculum and scleral ROP and non-ROP infants are shown in Table 3. Infants with ROP depressors. Before examination, proper dilation of the pupil was performed had significantly higher average and maximum glucose values. using a topical anesthetic followed by an eye drop combination Hypoglycemia was detected in 13 (20%) of the studied population. (0.2% cyclopentolate and 1% phenylephrine). Stages of severity and the The incidence of ROP in infants with hypoglycemia was 15%, zones for the extent of ROP were categorized by the lowest zone whereas ROP in non-hypoglycemic patients was 33%; this differ- and the highest stage observed in each eye using the international ence was not significant (P = 0.315). classification of ROP.33,34 Duration of respiratory support (continuous positive airway pressure or mechanical ventilation), duration of oxygen therapy Glucose measurement and the concentration of the administered oxygen (FiO2)duringthe Routine laboratory investigations were obtained from all cases, including first 3 days were all significantly higher in ROP cases (Table 4). complete blood picture, blood chemistry and blood gases. In addition, Compared with non-ROP infants, infants who developed ROP had glucose concentration was checked using point-of-care glucometer delayed onset to introduce enteral feeds (7.6 ± 4.5 vs 5 ± 3.4 days, (Optuim Xceed, Abbott, lake Forest, IL, USA).