Retinopathy of prematurity (ROP) is a disorder marked by abnormal neurovascular development in premature infants.1 Although mild forms of ROP may spontaneously resolve, severe cases may result in retinal scarring, detachment, and blindness if untreated.2 Risk factors for developing ROP generally belong to two categories: prenatal and postnatal factors. Prenatal factors, which influence the vulnerability of the retina due to lack of vascular development prior to birth, include low gestational age (GA) and very low birth weight (VLBW).3 Perinatal factors associated with ROP include oxygen supplementation, maternal chorioamnionitis, and decreased levels of insulin-like growth factor I (IGF-I).4
Short-term postnatal weight changes have been used to develop screening algorithms for the early detection of severe ROP in very preterm infants. The Weight IGF-I Neonatal ROP (WINROP) algorithm was developed to identify preterm infants at greatest risk for ROP based on GA, BW, and weekly weight measurements.4 Poor postnatal weight gain has been found to be highly predictive of developing severe ROP.5 In a cohort of 318 preterm infants in the United States, the WINROP algorithm identified all 28 infants who developed severe ROP at a median of 9 weeks prior to diagnosis, with a specificity of 82%.6 The Colorado–Retinopathy of Prematurity model (CO-ROP) incorporates postnatal weight gain in the first month of life to identify infants at risk of both high-grade and low-grade ROP and demonstrated a sensitivity of 96.4% for detecting ROP.7
Our study aimed to examine the relationship between postnatal weight gain and development of ROP among very preterm infants in southwestern Ontario. We specifically examined weight changes in the early postnatal period from birth to the first day of established full enteral feeding (FEF) and the subsequent period from FEF to hospital discharge. These two periods have different trajectories of daily weight changes. We hypothesized that poorer weight gain and longer durations from birth to FEF and from FEF to discharge/transfer would be associated with increased risk of ROP.
Patients and Methods
The health records of preterm infants admitted to the neonatal intensive care units (NICUs) at St. Joseph's Hospital and Victoria Hospital in London, Ontario, Canada, born between January 1, 2008, and June 1, 2015, with a BW of less than 1,250 g or a GA of less than 31 weeks were reviewed. A total of 456 infants met these criteria. Twenty-five infants with missing weight or nutritional data or who died before eye screening for ROP were excluded (Figure 1), resulting in a total of 431 infants included in the analysis. The cohort represented all of the very preterm infants born in southwestern Ontario, Canada, who were exclusively admitted to the two tertiary NICUs in the respective time periods. Data on neonatal and maternal characteristics at birth, morbidities, weight gain/loss, nutritional intake, and ROP screening results were collected. ROP screening was performed by three staff ophthalmologists, using indirect ophthalmoscopy with a 28/20 diopter lens, indentor, and eyelid speculum. Neonatal characteristics at birth that were collected included sex, GA, BW, birth length, head circumference, Apgar score at 5 minutes, and umbilical cord arterial pH. Maternal characteristics that were collected included age, gravidity, number of gestations, birth order, antenatal steroid therapy, diabetes mellitus, antepartum hemorrhage, chorioamnionitis, hypertension, mode of delivery, oligohydramnios/polyhydramnios, and premature rupture of membranes. Infant clinical data collected included respiratory distress syndrome, pulmonary interstitial emphysema, bronchopulmonary dysplasia (BPD), endotracheal tube infection, urinary tract infection, cerebrospinal fluid infection, sepsis, patent ductus arteriosus (PDA), medical or surgical treatment for PDA, intraventricular hemorrhage, periventricular leukomalacia, necrotizing enterocolitis, blood transfusions, and surfactant administration. This study was approved by Western University's Research Ethics Board and adhered to the tenets of the Declaration of Helsinki.
Flow diagram of the study population. NICU = neonatal intensive care unit; BW = birth weight; GA = gestational age; ROP = retinopathy of prematurity.
Chi-square and t tests were used to compare infants with and without ROP (including stages 1 to 5) for categorical and continuous variables, respectively. Weight was collected at birth, days 7, 14, 21, 28, and 42 of life, FEF, and discharge/transfer. Weight changes between each consecutive time point from birth to day 42 of life were calculated and expressed as weight gain per day (g/day) for each period. Multivariable logistic regression analysis was used to compare time durations and weight velocities between infants with and without ROP while controlling for confounding variables, including GA and BW, which previous studies have already shown to be negatively correlated with ROP.3 Variables significant at the bivariate level were entered into the regression model and subsequently removed at a P value of less than .05 in a backward elimination strategy. SPSS software (version 24; IBM Corporation, Armonk, NY) was used for all analyses, and P values of less than .05 were considered statistically significant.
In total, 431 infants met inclusion criteria and 179 (42%) were diagnosed as having ROP (Figure 1). The mean GA of infants with ROP was 26.3 ± 0.1 weeks, which was significantly lower than the mean GA of infants without ROP of 28.0 ± 0.1 weeks (P < .001). Infants with ROP had significantly lower BW (851 ± 13 g) than infants without ROP (1,005 ± 11 g) (P < .001). Infants with ROP had a significantly higher mean birth weight standard deviation score (BWSDS) (standardized BW z-scores calculated using the 2013 Fenton Preterm Growth Chart8) of −0.01 ± 0.06 than infants without ROP (−0.17 ± 0.05) (P < .05).
Table 1 shows bivariate results of neonatal characteristics and risk factors that were statistically significant between infants with and without ROP. Comorbidities of ROP, including BPD, endotracheal tube infection, and blood transfusion, were more common among infants with ROP (P < .001). Diagnosis of PDA, medication for PDA, and surgical ligation for PDA were also more common among infants with ROP (P < .001).
Bivariate Analysis of Risk Factors Between Infants With and Without ROP
Figure 2 illustrates mean weights and postmenstrual ages of infants with and without ROP at birth, FEF, and discharge/transfer. Patients with ROP had a longer duration from birth to FEF (34.0 ± 2.0 vs 20.2 ± 1.0 days) and from FEF to discharge/transfer (62.1 ± 2.1 vs 44.1 ± 1.5 days) compared to patients without ROP (P < .001). Even after controlling for GA and BW, these differences remain significant (P < .05). The weight velocity from birth to FEF in patients with ROP was higher than in patients without ROP (10.1 ± 0.6 vs 5.5 ± 0.6 g/day, P < .001). However, the difference was no longer significant after controlling for GA, BW, and duration from birth to FEF, supporting the notion that the weight velocity is a function of time and is dependent on postnatal age. Likewise, there was no difference in the weight velocity from FEF to discharge/transfer when controlled for GA, BW, and duration from FEF to discharge/transfer.
Mean weight (g) ± standard error vs postmenstrual age (weeks) for infants with and without retinopathy of prematurity (ROP) at birth, first day of full enteral feeding (FEF), and discharge/transfer.
Patients with ROP had a lower weight velocity from day 7 to day 28 (17.5 ± 0.5 vs 20.5 ± 0.4 g/day, P < .001) than patients without ROP. However, the difference was no longer significant after controlling for GA and BW (P > .05). This supports the contention that both GA and BW are correlated with weight velocities.
The weight velocities and time durations assessed were also included in a multivariable logistic regression analysis model (Table 2) controlling for the three strongest predictors of ROP: GA, BPD, and surgical ligation for PDA. Duration from FEF to discharge/transfer remained associated with ROP in this model (P < .05, Table 2). Weight velocity from day 7 to day 28, weight velocity from birth to FEF, and duration from birth to FEF did not arise as independent risk factors for ROP.
Adjusted Weight Velocities and Time Durations in a Multivariable Logistic Regression Model Predicting ROP
Our study shows that for very preterm infants born in southwestern Ontario, postnatal weight velocities are not independent predictors for the development of ROP. Although current Canadian screening guidelines for ROP take only BW and GA into consideration,9 some studies report poor relative10–13 and absolute13,14 postnatal weight gain as risk factors for developing ROP. Screening algorithms for ROP, such as WINROP4–6 and CO-ROP,7 use poor post-natal weight gain as a predictor to identify infants at higher risk of developing ROP. Additionally, oxygen supplementation and comorbidities such as intraventricular hemorrhage, PDA, and BPD have also been identified as risk factors associated with ROP.15–18
In the current study, low BW was associated with the development of ROP. Infants with ROP had lower weight velocity from day 7 to day 28 than infants without ROP. However, the difference was no longer significant after controlling for GA and BW. Thus, it appears that poor weight velocity from day 7 to day 28 is associated with ROP, but it is not an independent risk factor due to its association with low GA and BW.
Infants with ROP also had higher weight velocity from birth to FEF. However, there was no significant difference in weight velocity from birth to FEF after controlling for GA, BW, and duration from birth to FEF. The higher weight velocity from birth to FEF in infants with ROP may be due to the longer time required to reach FEF, allowing for a longer period of accelerating growth following the weight nadir at approximately day 7 of life.
Longer duration from FEF to discharge/transfer was associated with ROP when controlling for GA and BW. This association remained significant in a multivariable logistic regression analysis model controlling for GA, BPD, and surgical ligation for PDA (Table 2). This result is consistent with administrative decisions for prolonging hospital stay because infants diagnosed as having ROP often have discharge/ transfer from the NICU deferred to facilitate follow-up retinal examinations.
Longer durations from birth to FEF were found to be associated with the development of ROP. Infants with ROP were found to require a longer time to reach FEF, even when controlling for GA and BW. However, duration from birth to FEF did not remain predictive of ROP in a multivariable logistic regression analysis model controlling for GA, BPD, and surgical ligation for PDA. These results suggest that difficulty in establishing enteral feeding, as indicated by a longer duration from birth to FEF, is a risk factor for ROP that is independent of BW and GA but strongly associated with comorbidities. Duration from birth to FEF may be regarded as a surrogate marker of the underlying biology that contributes to the risk of ROP and comorbidities in VLBW and very preterm infants. It may be important to include the time to FEF in screening tools for the development of ROP in very preterm infants. Such a screening algorithm would need to be tested in a prospective study.
Our study is limited in that certain known risk factors for the development of ROP, including the use of oxygen therapy, were not considered in the analysis. Factors affecting postnatal weight gain, including IGF-I, and parenteral and enteral nutrition, were also not measured. Although the severity of ROP is an important determinant of disease management and clinical outcomes,2 the severity of ROP was not assessed in the current study.
Our study found the duration from birth to the establishment of FEF in very preterm infants may be an additional risk factor for the development of severe ROP. Used in conjunction with BW and GA, the duration from birth to FEF may improve the reliability and efficiency of current screening tools for ROP. However, this was highly associated with the three strongest predictors of ROP: GA, BPD, and surgical ligation for PDA. Duration from birth to FEF may thus represent a surrogate marker that correlates to ROP and comorbidities in very pre-term and VLBW infants. Future studies may assess how FEF may be achieved more quickly in preterm infants, and whether earlier achievement of FEF in preterm infants reduces the risk of developing ROP.
- Hellström A, Smith LEH, Dammann O. Retinopathy of prematurity. Lancet. 2013;382:1445–1457. doi:10.1016/S0140-6736(13)60178-6 [CrossRef]
- Ashton N, Ward B, Serpell G. Effect of oxygen on developing retinal vessels with particular reference to the problem of retrolental fibroplasia. Br J Ophthalmol. 1954;38:397–432. doi:10.1136/bjo.38.7.397 [CrossRef]
- Wikstrand MH, Hård AL, Niklasson A, Smith L, Löfqvist C, Hellström A. Maternal and neonatal factors associated with poor early weight gain and later retinopathy of prematurity. Acta Paediatr. 2011;100:1528–1533. doi:10.1111/j.1651-2227.2011.02394.x [CrossRef]
- Löfqvist C, Hansen-Pupp I, Andersson E, et al. Validation of a new retinopathy of prematurity screening method monitoring longitudinal postnatal weight and insulin-like growth factor I. Arch Ophthalmol. 2009;127:622–627. doi:10.1001/archophthalmol.2009.69 [CrossRef]
- Lundgren P, Stoltz Sjöström E, Domellöf M, et al. WINROP identifies severe retinopathy of prematurity at an early stage in a nation-based cohort of extremely preterm infants. PLoS One. 2013;8:e73256. doi:10.1371/journal.pone.0073256 [CrossRef]
- Wu C, VanderVeen DK, Hellström A, Löfqvist C, Smith LEH. Longitudinal postnatal weight measurements for the prediction of retinopathy of prematurity. Arch Ophthalmol. 2010;128:443–447. doi:10.1001/archophthalmol.2010.31 [CrossRef]
- Cao JH, Wagner BD, McCourt EA, et al. The Colorado-retinopathy of prematurity model (CO-ROP): postnatal weight gain screening algorithm. JAAPOS. 2016;20:19–24.
- Fenton TR, Kim JH. A systematic review and meta-analysis to revise the Fenton growth chart for preterm infants. BMC Pediatr. 2013;13:59. doi:10.1186/1471-2431-13-59 [CrossRef]
- Jefferies A. Retinopathy of prematurity: recommendations for screening. Paediatr Child Health. 2010;15:667–670. doi:10.1093/pch/15.10.667 [CrossRef]
- Wallace DK, Kylstra JA, Phillips SJ, Hall JG. Poor postnatal weight gain: a risk factor for severe retinopathy of prematurity. JAAPOS. 2000;4:343–347.
- Fortes Filho JB, Bonomo PP, Maia M, Procianoy RS. Weight gain measured at 6 weeks after birth as a predictor for severe retinopathy of prematurity: study with 317 very low birth weight preterm babies. Graefes Arch Clin Exp Ophthalmol. 2009;247:831–836. doi:10.1007/s00417-008-1012-3 [CrossRef]
- Aydemir O, Sarikabadayi YU, Aydemir C, et al. Adjusted poor weight gain for birth weight and gestational age as a predictor of severe ROP in VLBW infants. Eye (Lond). 2011;25:725–729. doi:10.1038/eye.2011.29 [CrossRef]
- Wang ZH, Gao PF, Bai H, Li YY. Postnatal weight gain in very low birth weight infants in Beijing and the risk of retinopathy of prematurity. Int J Ophthalmol. 2015;8:1207–1210.
- Allegaert K, Vanhole C, Casteels I, et al. Perinatal growth characteristics and associated risk of developing threshold retinopathy of prematurity. JAAPOS. 2003;7:34–37.
- Zepeda-Romero LC, Lundgren P, Gutierrez-Padilla JA, et al. Oxygen monitoring reduces the risk for retinopathy of prematurity in a Mexican population. Neonatology. 2016;110:135–140. doi:10.1159/000445040 [CrossRef]
- Fortes Filho JB, Eckert GU, Procianoy L, Barros CK, Procianoy RS. Incidence and risk factors for retinopathy of prematurity in very low and in extremely low birth weight infants in a unit-based approach in southern Brazil. Eye (Lond). 2009;23:25–30. doi:10.1038/sj.eye.6702924 [CrossRef]
- Waheeb S, Alshehri K. Incidence of retinopathy of prematurity at two tertiary centers in Jeddah, Saudi Arabia. Saudi J Ophthalmol. 2016;30:109–112. doi:10.1016/j.sjopt.2016.02.006 [CrossRef]
- Gebesce A, Uslu H, Keles E, et al. Retinopathy of prematurity: incidence, risk factors, and evaluation of screening criteria. Turk J Med Sci. 2016;46:315–320. doi:10.3906/sag-1407-127 [CrossRef]
Bivariate Analysis of Risk Factors Between Infants With and Without ROP
|Parameter||ROP (n = 179)||No ROP (n = 252)||P|
|Gestational age (wks), mean ± SEM||26.3 ± 0.1||28.1 ± 0.1||< .001a|
|Male, number (%)||95 (53)||124 (49)||.429|
|Birth weight (g), mean ± SEM||851 ± 13||1,005 ± 11||< .001a|
|Birth weight standard deviation score, mean ± SEM||−0.01 ± 0.06||−0.17 ± 0.05||.037b|
|Apgar at 5 minutes, median||6||7||.003c|
|Cord blood arterial pH, mean ± SEM||7.261 ± 0.008||7.236 ± 0.007||.024b|
|Gestational hypertension, number (%)||24 (13)||58 (23)||.012b|
|Vaginal birth, number (%)||90 (50)||94 (37)||.007c|
|Bronchopulmonary dysplasia, number (%)||128 (72)||75 (30)||< .001a|
|Endotracheal tube infection, number (%)||90 (50)||50 (20)||< .001a|
|Sepsis, number (%)||58 (32)||47 (19)||.001c|
|Patent ductus anteriosus, number (%)||131 (73)||122 (48)||< .001a|
|Medication for PDA, number (%)||108 (60)||89 (35)||< .001a|
|Surgical ligation for PDA, number (%)||36 (20)||9 (4)||< .001a|
|Intraventricular hemorage, number (%)||91 (51)||88 (35)||.001c|
|Necrotizing enterocolitis, number (%)||27 (15)||17 (7)||.005c|
|Blood transfusion, number (%)||142 (79)||122 (48)||< .001a|
|Day 7 to day 28 weight velocity (g/day), mean ± SEM||17.5 ± 0.5||20.5 ± 0.4||< .001a|
|Birth-FEF weight velocity (g/day), mean ± SEM||10.1 ± 0.6||5.5 ± 0.6||< .001a|
|Birth-FEF duration (days), mean ± SEM||34.0 ± 2.0||20.2 ± 1.0||< .001a|
|FEF-discharge duration (days), mean ± SEM||62.1 ± 2.1||44.1 ± 1.5||< .001a|
Adjusted Weight Velocities and Time Durations in a Multivariable Logistic Regression Model Predicting ROP
|Parameter||Regression Coefficient (SE)||Adjusted Odds Ratio||P|
|Day 7 to day 28 weight velocity (g/day), mean ± SE||0.027 ± 0.022||1.027||.216|
|Birth-FEF weight velocity (g/day), mean ± SE||0.027 ± 0.015||1.027||.07|
|Birth-FEF duration (days), mean ± SE||0.008 ± 0.007||1.008||.264|
|FEF-discharge duration (days), mean ± SE||0.013 ± 0.005||1.013||.018a|