- Email: [email protected]
Postnatal corticosteroids and risk of retinopathy of prematurity
Accepted Manuscript Postnatal corticosteroids and risk of retinopathy of prematurity Tammy Z. Movsas, MD, MPH, Alan R. Spitzer, MD, Ira H. Gewolb, MD PII:
S1091-8531(16)30107-0
DOI:
10.1016/j.jaapos.2016.05.008
Reference:
YMPA 2424
To appear in:
Journal of AAPOS
Received Date: 11 February 2016 Revised Date:
19 May 2016
Accepted Date: 28 May 2016
Please cite this article as: Movsas TZ, Spitzer AR, Gewolb IH, Postnatal corticosteroids and risk of retinopathy of prematurity, Journal of AAPOS (2016), doi: 10.1016/j.jaapos.2016.05.008. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Postnatal corticosteroids and risk of retinopathy of prematurity Tammy Z. Movsas, MD, MPH,a Alan R. Spitzer, MD,b and Ira H. Gewolb, MDc
RI PT
Author affiliations: aZietchick Research Institute, Plymouth, Michigan and Midland County Department of Public Health, Midland, Michigan; bMEDNAX Services-Pediatrix Medical Group, Sunrise, Florida; cDepartment of Pediatrics & Human Development, Michigan State University, East Lansing, Michigan Submitted February 11, 2016. Revision accepted May 28, 2016.
SC
Correspondence: Tammy Z. Movsas, MD, MPH, Medical Director, Midland County Dept. of Public Health, 220 West Ellsworth St, Midland, MI 48640 (email: [email protected]).
AC C
EP
TE D
M AN U
Word count: 2,666 Abstract only: 140
ACCEPTED MANUSCRIPT
Abstract Purpose To investigate the association between postnatal steroids and retinopathy of prematurity (ROP)
RI PT
in neonates born with birth weights at the limit of viability (<500 g). Methods
Data from the Pediatrix BabySteps Clinical Warehouse was retrospectively reviewed. The study
SC
population consisted of 1,472 neonates with birth weights of <500 g who were discharged alive from 167 NICUs between 1996 and 2013. Statistical significance for unadjusted comparisons
M AN U
between groups was determined using the χ2 or t test. Logistic regression was used to calculate odds of ROP. Results
In multivariate analysis, the odds of any ROP for steroid treated infants was 1.6 (95% CI, 1.2-
Conclusions
TE D
2.2) compared to nontreated infants; the odds of advanced ROP was 1.7 (95% CI, 1.3-2.3).
In our large study cohort of critically low birth weight infants ROP was more common in
AC C
EP
neonates exposed to postnatal steroids.
ACCEPTED MANUSCRIPT
Because of the beneficial effect of corticosteroids on lung function, especially in infants who are ventilator dependent, corticosteroids are frequently administered to very low birth weight (VLBW) neonates (birth weight <1500 g) to treat established or evolving lung disease.1
RI PT
However, over the last several years, there has been increasing concern that adverse
neurodevelopmental effects may result from postnatal steroid use. A Cochrane Database System Review examined randomized controlled trials that had investigated the long-term outcome of
SC
preterm infants who had been treated with postnatal steroids within the first week of life: 12 trials reported an increase in adverse neurological effects including developmental delay,
M AN U
cerebral palsy, and abnormal neurological examination in the infants who had been treated with the postnatal steroids.2 Because of the potential for adverse neurological effects, the American Academy of Pediatrics has advised that postnatal steroids be limited to certain clinical conditions.3 However, the debate regarding the benefit of postnatal corticosteroid as a respiratory
continues.
TE D
rescue therapy versus the potential harm of neurodevelopmental impairment in VLBW neonates
In addition to potential neurodevelopmental impairment from steroids, several
EP
observational studies of VLBW neonates have reported a significant association between postnatal steroid administration and an increased risk of retinopathy of prematurity (ROP).1,4-5
AC C
However, many clinicians have been skeptical of these results because multivariate analyses cannot adjust for the marked clinical differences between VLBW neonates who receive steroids versus those who do not. This problem occurs because VLBW neonates represent a heterogeneous group of infants who range from those of very immature gestational age to those who are more mature but extremely growth retarded.6 In addition, many substantive differences in risk factors for ROP exist in chronically ventilated neonates compared to premature infants
ACCEPTED MANUSCRIPT
without significant pulmonary disease. The purpose of this study was to investigate the association between postnatal steroids and ROP risk by examining ROP risk in a large cohort of preterm neonates born at birth weights
RI PT
at the limit of viability (<500 g). This cohort represents a more homogeneous set of neonates because neonates at these critically low birth weights (and gestational ages) are at the highest vulnerability for a host of neonatal morbidities including ROP and bronchopulmonary dysplasia.
SC
Thus, clinical differences between steroid-treated and untreated neonates can be minimized. We utilized data from Pediatrix BabyStep Clinical Data Warehouse, one of the world’s
M AN U
largest repositories of neonatal data, to examine the incidence of ROP in a large cohort of preterm infants with birthweight of <500 g. This database includes information on >1,120,000 infants who had been cared for by Pediatrix Medical Group providers at US hospitals since 1996.7-8 Data is automatically extracted from the daily medical record notes on each child
TE D
admitted by Pediatrix physicians At the current time, the Pediatrix Medical Group cares for nearly 25% of neonates in more than 300 NICUs across the U.S. The Pediatrix BabySteps Clinical Data Warehouse captures diagnostic and procedure coding data from the Pediatrix
EP
Medical Group electronic health record, which strictly adheres to the American Academy of Pediatrics Perinatal Section Coding Committee Guidelines.7
AC C
Methods
All data entry for the Pediatrix BabySteps Clinical Data Warehouse is verified on an ongoing basis by the Pediatrix Clinical Data Warehouse Information Technology team and by a database manager. The data from the Pediatrix BabySteps Clinical Data Warehouse is annually certified as deidentified and are compliant with the US Health Insurance Portability and Accountability Act of 1996. The data has been approved for use in research studies by the Western Institutional
ACCEPTED MANUSCRIPT
Review Board. In addition, an exemption for any additional institutional review board approval for use of the deidentified dataset in this study was granted by the Michigan State University Institutional Review Board.
RI PT
The following 3 inclusion criteria were used for this study population were: (1) birth weight <500 g; (2) hospitalization survival (ie, discharged from hospital alive); (3) ophthalmic ROP examination results available.
SC
The diagnosis and stage of ROP in the medical records was assigned by the board-
certified ophthalmologists who performed the retinal ROP examinations. The retinal findings
M AN U
were standardized by classifying the examination results according to the International Classification of Retinopathy of Prematurity.9-11 In most cases, multiple ROP examinations had been performed for each infant; the most advanced stage of ROP noted in the medical record was recorded as the infant’s ROP stage. Although infants may have received additional ROP follow-
TE D
up after hospital discharge, no post-discharge examination results were available for this study. Given that the average age at discharge was at a postmenstrual age close to term, however, it is highly probable that ROP severity would have reached its peak for the vast majority of the study
EP
participants before hospital discharge.
Study participants, born between 1996 and 2013, had been NICU inpatients at 167
AC C
different hospitals across the United States. Our study dataset included binary data (yes/no) on steroid exposure during pregnancy (ie, antenatal steroid) and after birth (postnatal steroid exposure). Steroid exposure included exposure to any type of steroids (betamethasone, dexamethasone, hydrocortisone, prednisolone, prednisone) for any indication. Also, the neonate may have been exposed to more than one type of steroid and for differing durations of treatment. Statistical significance for unadjusted comparisons between groups was determined with
ACCEPTED MANUSCRIPT
Pearson’s χ2 test or t test (for parametric comparison of means) or Wilcoxon rank sum test (for nonparametric comparison of median Apgar scores). Multivariate models were adjusted for several covariates. To adjust for secular trends in steroid use over the years, different steroid
RI PT
practices in different hospitals and differences in ROP assessment by different ophthalmologists, year of birth, transfer status (born in NICU hospital or transferred in from another hospital), and NICU hospital facility were included as covariates. Because gestational age, birthweights and
SC
oxygen exposure/ventilation (number of days FiO2 >0.21, days on high-flow nasal cannula, days on ventilation, and days on continuous positive airway pressure [CPAP]) are key risk factors for
M AN U
ROP, these were included as covariates. Comorbidities and other demographics that significantly differed between the steroid treated and untreated groups were included as covariates (patent ductus arteriosus, brain hemorrhage, necrotizing enterocolitis, bronchopulmonary dysplasia, sepsis, Apgar scores, and chronological age at hospital discharge). Antenatal steroid exposure
Results
TE D
and sex were also included as covariates.
Data from 1,472 infant entries from the Pediatrix BabySteps Clinical Data Warehouse repository
EP
satisfied all three inclusion criteria and were used to compile the study dataset. Table 1 compares the characteristics of the postnatal steroid-treated and untreated neonates. Of the 1,472 infants,
AC C
1059 (71.9%) received postnatal corticosteroids and 413 (28%) did not. There were no significant differences between groups with regard to race, sex, or antenatal steroid exposure (ie, steroid exposure during fetal life). There were small but statistically significant differences between groups in mean birthweight, mean gestational age, and median Apgar scores. As expected, there were large, significant differences between groups in variables relating to oxygen exposure.
ACCEPTED MANUSCRIPT
Table 2 provides the unadjusted incidence of ROP and other comorbidities of steroidtreated and nontreated neonates. The overall incidence of ROP (of any stage) for the entire cohort was 76.6% (1128/1472), and the overall incidence of advanced stage ROP (stages 3, 4, or
RI PT
5) was 31.3%. The incidence of any ROP was significantly higher (P < 0.0001) in steroid-treated infants (80.5%) than in nontreated infants (66.8%); the incidence of advanced-stage ROP was also significantly higher (P < 0.0001) in the former (35.3%) compared to the latter group
SC
(21.1%). Steroid-treated infants also had a significantly higher incidence of bronchopulmonary dysplasia, sepsis, patent ductus arteriosus, and intracranial hemorrhage compared to nonsteroid
M AN U
treated neonates.
Table 3 provides the univariate and multivariate effect of postnatal steroid use and sepsis on the odds of ROP incidence for any ROP and for advanced ROP. All multivariate logistic regression included the covariates of postnatal steroids, bronchopulmonary dysplasia, and sepsis
TE D
as well as the covariates of hospital facility, transfer status, antenatal steroid use, birth year, gestational age, birthweight, patent ductus arteriosus, Apgar scores, brain hemorrhage, discharge age, and sex as well as each type of O2 exposure (number of days on FiO2, artificial ventilation,
EP
high flow nasal cannula, and CPAP).
In multivariate analysis there was a 1.6 times greater odds (95% CI, 1.2-2.2) of any ROP
AC C
for steroid-treated infants compared to nontreated infants and 1.7 times greater odds (95% CI, 1.3-2.3) of advanced ROP for steroid-treated infants. In multivariate analysis the odds were 1.5 times greater of any ROP for infants with sepsis (95% CI, 1.1-2.0) and 1.4 times greaterof advanced ROP (95% CI, 1.1-1.8). Discussion The current study indicates that there is a higher risk of ROP in steroid-treated infants than in
ACCEPTED MANUSCRIPT
those infants not treated with steroids. The association between postnatal steroid use in VLBW infants and ROP has been long debated. Some randomized controlled trials involving postnatal steroid show trends toward increased ROP risk but few show any significant difference.1-2 In
RI PT
general, the postnatal steroid randomized controlled trials assess ROP as a secondary outcome and are not primarily designed to detect ROP risk. For example, most randomized controlled trials do not stratify birth weight below 1500 g and there are marked differences between infants
SC
born at different gestational ages in this group overall.1-2 As a result, ROP risk analysis is
within the overall VLBW population.12
M AN U
problematic since birthweight persists as a fundamental, independent ROP risk factor, even
In addition to the lack of birth weight stratification, diagnostic ROP protocols in some of the randomized control trials veer from accepted standards.13 For example, Suske and colleagues14 assessed ROP via a single retinal examination performed at 4 weeks’ postpartum
TE D
age. Because ROP can occur and progress anytime until retinal maturation,13 their method would likely underestimate both ROP incidence and severity in any study population. Another randomized controlled study assessed for “acute ROP” but offered no further definition for this
EP
diagnosis.15
Two separate recent Cochrane reviews performed meta-analyses (integrating data from
AC C
several small randomized controlled trials to examine the effect of the timing of steroid initiation on risk of various outcomes (including ROP) for preterm infants.1-2 These reviews reported that late steroids (ie, initiated after 1 week of life) significantly increased ROP risk, whereas early steroids (ie, initiated after birth but before one of life) significantly decrease risk.1-2 Because the upper limit of their 95% confidence interval limit for early steroid meta-analysis bordered on 1 (0.97 to be exact), 2 the Cochrane meta-analysis for the effect of early initiation of steroids on
ACCEPTED MANUSCRIPT
ROP risk should be interpreted with caution. In addition, the Cochrane meta-analyses incorporated data from some aforementioned randomized controlled studies with substandard ROP methodology that added additional concerns about the validity of their conclusions with
RI PT
respect to ROP risk.14-15 Thus, to date, the available meta-analyses do not provide convincing evidence that postnatal steroids are either harmful or beneficial with respect to ROP.
Given these issues with the postnatal steroid randomized controlled trials and meta-
SC
analyses, the present study undertook a large retrospective review in order to shed light on the ROP-steroid association. We specifically chose our study population to consist of neonates with
M AN U
birthweight of <500 g because this category of neonates is generally more homogeneous than those that include larger birth weight infants; thus, clinical differences between treated/untreated groups, though not entirely eliminated, are minimized. In addition, our study population was limited to infants who survived hospitalization in order to eliminate death as a competing
TE D
outcome with ROP. In other words, because all infants survived, all had an “equal opportunity” for ROP to have either developed or progressed. However, because survival was an inclusion criteria, our study population was female-weighted (male sex is known to be a strong risk factor
EP
for death in extremely premature infants).16 For this reason, sex was included as a covariate in our multivariate analyses.
AC C
Given that bronchopulmonary dysplasia is one of the main indications for corticosteroids,
we found large expected differences between our groups (treated/untreated) with respect to oxygen administration. We adjusted for oxygen exposure in our multivariate analyses. It is important to remember, however, that oxygen exposure per se is not the key risk factor in the development of ROP; rather, tissue oxygen content (paO2 or blood saturation) is the critical element. Although oxygen exposure was used as a marker for hyperoxemia, it is important to
ACCEPTED MANUSCRIPT
note that the two are not always the same. Despite statistically significant differences in birth weight, gestational age, and other characteristics/morbidities between the treated and untreated groups, the magnitude of the
RI PT
differences between groups was generally small; nevertheless, they were adjusted for in
multivariate analysis. Of note, some aspects of the International Classification of Retinopathy of Prematurity (ICROP) were updated by a consensus statement published in 2005.11 Thus, study
SC
participants born before 2005 were staged by the ICROP of 1987 and those born after 2005 were staged according to the updated ICROP. That said, the modifications to the original ICROP were
M AN U
relatively few and did not affect whether or not an infant would be diagnosed with ROP and also did not affect whether or not the ROP would have been classified as early versus late stage. Thus, our results are not affected by this change. Nonetheless, we adjusted for year of birth in multivariate analyses.
TE D
Despite equal medians of Apgar scores at 1 minute in the treated and untreated group, a statistical difference (P < 0.0001) between these groups was detected for this variable by the Wilcoxon rank sum test (ie, lower scores below median for steroid treated group compared to
EP
non-treated group) This finding can occur because the Wilcoxon test ranks all of the observations from both groups and then sums the ranks from one of the groups and compares this to the
test.
AC C
expected rank sum. In other words, the nonparametric Wilcoxon test is not specifically a median
With multivariate analysis, the use of postnatal steroids was a significant ROP risk factor
for critically low birth weight newborns. It yielded between 1.6 and 1.7 times an increase in the odds of ROP. In neonates of birth weight <1500 g, ROP incidence approximates 25% and in neonates with birth weight <1000 g, ROP incidence exceeds 50%.17-18 Given this high incidence
ACCEPTED MANUSCRIPT
of ROP in the VLBW population, even a modest increase in the odds of ROP puts a large number of additional VLBW infants at risk for lifelong visual disability. We identified sepsis to be another significant ROP risk factor for any ROP as well as for
RI PT
advanced ROP. Sepsis has been correlated with ROP in other VLBW studies as well.19This finding is relevant because the aforementioned Cochrane meta-analysis showed a trend toward increased sepsis in steroid-treated neonates1; thus, adverse effects of steroids, such as increase in
SC
sepsis, may additionally raise ROP risk.
The retrospective nature of this study prevents drawing any causal inferences. In addition,
M AN U
the binary nature of the steroid data does not permit for the evaluation of what steroid type, mode of delivery, timing, duration, or dosage are associated with ROP risk. Despite these limitations, this study of a large database of critically low birth weight survivors indicates that steroid-treated infants have a modest but significantly increased risk for ROP. This observation is of potential
TE D
clinical significance, because children with a history of ROP are not only at increased risk for visual impairments from the ROP itself but are also at increased risk for developing other ocular
AC C
EP
disorders later in life.20-23
ACCEPTED MANUSCRIPT
References 1.
Doyle LW, Ehrenkranz RA, Halliday HL. Late (>7 days) postnatal corticosteroids for chronic lung disease in preterm infants. Cochrane Database Syst Rev 2014;5:CD001145. Doyle LW, Ehrenkranz RA, Halliday HL. Early (<8 days) postnatal corticosteroids for
RI PT
2.
preventing chronic lung disease in preterm infants. Cochrane Database Syst Rev 2014;5:CD001146.
Watterberg KL. Policy statement—postnatal corticosteroids to prevent or treat
SC
3.
bronchopulmonary dysplasia. Pediatrics 2010;126:800-808.
Batton DG, Roberts C, Trese M, Maisels MJ. Severe retinopathy of prematurity and
M AN U
4.
steroid exposure. Pediatrics 1992;90:534-6. 5.
Ramanathan R, Siassi B, deLemos RA. Severe retinopathy of prematurity in extremely low birth weight infants after short-term dexamethasone therapy. J Perinatol
6.
TE D
1995;15:178-82; quiz 183-74.
Arnold CC, Kramer MS, Hobbs CA, McLean FH, Usher RH. Very low birth weight: a problematic cohort for epidemiologic studies of very small or immature neonates. Am J
7.
EP
Epidemiol 1991;134:604-13.
Spitzer AR, Ellsbury D, Clark RH. The Pediatrix BabySteps(R) Data Warehouse—a
AC C
unique national resource for improving outcomes for neonates. Indian J Pediatr 2015;82:71-9.
8.
Olsen IE, Groveman SA, Lawson ML, Clark RH, Zemel BS. New intrauterine growth curves based on United States data. Pediatrics 2010;125:e214-24.
9.
Committee for the Classification of Retinopathy of Prematurity. An International Classification of Retinopathy of Prematurity. Arch Ophthalmol 1984;102:1130-34.
ACCEPTED MANUSCRIPT
10.
International Committee for the Classification of the Late Stages of Retinopathy of Prematurity. An International Classification of Retinopathy of Prematurity. II. The classification of retinal detachment. Arch Ophthalmol 1987;105:906-12. The International Classification of Retinopathy of Prematurity revisited. Arch
RI PT
11.
Ophthalmol 2005;123:991-9. 12.
Miller MM, Revenis ME, Lai MM, et al. Risk and clinical course of retinopathy of
Fierson WM. Screening examination of premature infants for retinopathy of prematurity. Pediatrics 2013;131:189-95.
14.
M AN U
13.
SC
prematurity in 78 infants of gestational age 22-25 weeks. J AAPOS 2014;18:266-70.
Suske G, Oestreich K, Varnholt V, Lasch P, Kachel W. Influence of early postnatal dexamethasone therapy on ventilator dependency in surfactant-substituted preterm infants. Acta Paediatr 1996;85:713-18.
Yeh TF, Torre JA, Rastogi A, Anyebuno MA, Pildes RS. Early postnatal dexamethasone
TE D
15.
therapy in premature infants with severe respiratory distress syndrome: a double-blind, controlled study. J Pediatr 1990;117(2 Pt 1):273-82. Tyson JE, Parikh NA, Langer J, Green C, Higgins RD. Intensive care for extreme
EP
16.
prematurity--moving beyond gestational age. N Engl J Med 2008;358:1672-81. Lermann VL, Fortes Filho JB, Procianoy RS. The prevalence of retinopathy of
AC C
17.
prematurity in very low birth weight newborn infants. J Pediatr (Rio J) 2006;82:27-32.
18.
Wheeler DT, Dobson V, Chiang MF, et al. Retinopathy of prematurity in infants weighing less than 500 grams at birth enrolled in the early treatment for retinopathy of prematurity study. Ophthalmology 2011;118:1145-51.
19.
Kavurt S, Ozcan B, Aydemir O, Bas AY, Demirel N. Risk of retinopathy of prematurity
ACCEPTED MANUSCRIPT
in small for gestational age premature infants. Indian Pediatr 2014;51:804-6.
20.
Hartnett ME, Gilbert MM, Hirose T, Richardson TM, Katsumi O. Glaucoma as a cause
RI PT
of poor vision in severe retinopathy of prematurity. Graefes Arch Clin Exp Ophthalmol 1993;231:433-8. 21.
Lee RW, Mayer EJ, Markham RH. The aetiology of paediatric rhegmatogenous retinal
22.
SC
detachment: 15 years experience. Eye (Lond) 2008;22:636-40.
Yang CS, Wang AG, Sung CS, Hsu WM, Lee FL, Lee SM. Long-term visual outcomes
years. Eye (Lond) 2010;24:14-20.
Krolicki TJ, Tasman W. Cataract extraction in adults with retinopathy of prematurity.
EP
TE D
Arch Ophthalmol 1995;113:173-77.
AC C
23.
M AN U
of laser-treated threshold retinopathy of prematurity: a study of refractive status at 7
ACCEPTED MANUSCRIPT
Table 1. Demographics and characteristics of postnatal steroid-treated and untreated neonates in critically low birth weight cohort (birth weight <500 g; N =1472) No steroids N = 413
>0.05a a >0.05
RI PT
134 (32.4) 157 (38.0) 135 (32.7) 12 (2.9) 109 (26.4) 311 (75.8) 4 25.2 ± 2.2 461.8 ± 40.4 14.8 ± 6.5 68.0 ± 53.9 14.2 ± 19.4
P value
a
>0.05 b <0.0001 c <0.0001 c <0.05 <0.001c c <0.0001 c <0.001
SC
Postnatal steroids-treated N = 1059 Sex: male (%) 415 (39.1) White 449 (42.4) Black 345 (32.6) 34 (3.2) Asian Other/Unknown 231 (21.9) +antenatal steroid exposure (%) 831 (78.6) Median 1-minute Apgar score 4 GA, weeks, mean ± SD 24.6 ± 1.5 BW, g, mean ± SD 456.6 ± 43.2 CA at discharge, weeks, mean ± SD 17.6 ± 6.6 No. days FiO2 >.21, mean ± SD 109.8 ± 52.5 No. days on high-flow nasal cannula, mean 18.3 ± 22.2 ± SD No. days on ventilation, mean ± SD 62.5 ± 37.9 No. days on CPAP, mean ± SD 17.0 ± 19.4
c
33.1 ± 33.0 <0.001 c 14.7 ± 19.9 <0.05
M AN U
Characteristic
BW, birth weight; CA, chronological age; CPAP, continuous positive airway pressure; GA, gestational age; SD, standard deviation. a
Chi squared. Wilcoxon rank square test. c t test.
AC C
EP
TE D
b
ACCEPTED MANUSCRIPT
Table 2. Comparison of the prevalence of ROP and other comorbidities in neonates of birth weight <500 g treated with postnatal steroids and not treated with steroids (N = 1,472)
852 (80.5) 374 (35.3)
276 (66.8) 87 (21.1)
<0.0001 <0.0001
122 (11.5) 735 (69.4) 615 (58.1) 95 (9.0) 55 (5.2) 782 (73.8) 402 (37.9)
39 (9.4) 227 (55.0) 185 (44.8) 29 (7.0) 15 (3.6) 203 (49.1) 107 (25.9)
>0.05 <0.0001 <0.0001 >0.05 >0.05 <0.001 <0.0001
M AN U
ROP, retinopathy of prematurity.
SC
ROP, n (%) Any stage Advanced (stage 3, 4, 5) Other disorders, n (%) Congenital heart disorders Patent ductus arteriosus Intracranial hemorrhage (any type) Necrotizing enterocolitis (medical) Necrotizing enterocolitis (surgical) Bronchopulmonary dysplasia Sepsis a 2
AC C
EP
TE D
χ test.
a
RI PT
Postnatal steroid- No postnatal P value treated steroids N = 1,059 N = 413
ACCEPTED MANUSCRIPT
Postnatal steroids Sepsis
Postnatal steroids Sepsis
Odds of ROP, any stage (95% CI) Univariate logistic Multivariate logistic a regression regression 2.0 (1.6-2.6) 1.6 (1.2-2.2) 2.0 (1.5-2.6) 1.5 (1.1-2.0) Odds of advanced ROP (95% CI) Univariate logistic Multivariate logistic a regression regression 2.0 (1.6-2.7) 1.7 (1.3-2.3) 1.8 (1.5-2.3) 1.4 (1.1-1.8)
CI, confidence interval; ROP, retinopathy of prematurity.
SC
a
RI PT
Table 3. Effect of postnatal steroids on odds of ROP incidence in critically low birthweight neonates (birth weight <500 g)
AC C
EP
TE D
M AN U
Multivariate logistic regressions included covariates of postnatal steroids, bronchopulmonary dysplasia, sepsis, hospital facility, transfer status, antenatal steroid use, birth year, gestational age, birthweight, patent ductus arteriosus, Apgar scores and each type of O2 exposure (no. days on FiO2, vent, high-flow nasal cannula, CPAP), brain hemorrhage, discharge age, and sex.
S1091-8531(16)30107-0
DOI:
10.1016/j.jaapos.2016.05.008
Reference:
YMPA 2424
To appear in:
Journal of AAPOS
Received Date: 11 February 2016 Revised Date:
19 May 2016
Accepted Date: 28 May 2016
Please cite this article as: Movsas TZ, Spitzer AR, Gewolb IH, Postnatal corticosteroids and risk of retinopathy of prematurity, Journal of AAPOS (2016), doi: 10.1016/j.jaapos.2016.05.008. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Postnatal corticosteroids and risk of retinopathy of prematurity Tammy Z. Movsas, MD, MPH,a Alan R. Spitzer, MD,b and Ira H. Gewolb, MDc
RI PT
Author affiliations: aZietchick Research Institute, Plymouth, Michigan and Midland County Department of Public Health, Midland, Michigan; bMEDNAX Services-Pediatrix Medical Group, Sunrise, Florida; cDepartment of Pediatrics & Human Development, Michigan State University, East Lansing, Michigan Submitted February 11, 2016. Revision accepted May 28, 2016.
SC
Correspondence: Tammy Z. Movsas, MD, MPH, Medical Director, Midland County Dept. of Public Health, 220 West Ellsworth St, Midland, MI 48640 (email: [email protected]).
AC C
EP
TE D
M AN U
Word count: 2,666 Abstract only: 140
ACCEPTED MANUSCRIPT
Abstract Purpose To investigate the association between postnatal steroids and retinopathy of prematurity (ROP)
RI PT
in neonates born with birth weights at the limit of viability (<500 g). Methods
Data from the Pediatrix BabySteps Clinical Warehouse was retrospectively reviewed. The study
SC
population consisted of 1,472 neonates with birth weights of <500 g who were discharged alive from 167 NICUs between 1996 and 2013. Statistical significance for unadjusted comparisons
M AN U
between groups was determined using the χ2 or t test. Logistic regression was used to calculate odds of ROP. Results
In multivariate analysis, the odds of any ROP for steroid treated infants was 1.6 (95% CI, 1.2-
Conclusions
TE D
2.2) compared to nontreated infants; the odds of advanced ROP was 1.7 (95% CI, 1.3-2.3).
In our large study cohort of critically low birth weight infants ROP was more common in
AC C
EP
neonates exposed to postnatal steroids.
ACCEPTED MANUSCRIPT
Because of the beneficial effect of corticosteroids on lung function, especially in infants who are ventilator dependent, corticosteroids are frequently administered to very low birth weight (VLBW) neonates (birth weight <1500 g) to treat established or evolving lung disease.1
RI PT
However, over the last several years, there has been increasing concern that adverse
neurodevelopmental effects may result from postnatal steroid use. A Cochrane Database System Review examined randomized controlled trials that had investigated the long-term outcome of
SC
preterm infants who had been treated with postnatal steroids within the first week of life: 12 trials reported an increase in adverse neurological effects including developmental delay,
M AN U
cerebral palsy, and abnormal neurological examination in the infants who had been treated with the postnatal steroids.2 Because of the potential for adverse neurological effects, the American Academy of Pediatrics has advised that postnatal steroids be limited to certain clinical conditions.3 However, the debate regarding the benefit of postnatal corticosteroid as a respiratory
continues.
TE D
rescue therapy versus the potential harm of neurodevelopmental impairment in VLBW neonates
In addition to potential neurodevelopmental impairment from steroids, several
EP
observational studies of VLBW neonates have reported a significant association between postnatal steroid administration and an increased risk of retinopathy of prematurity (ROP).1,4-5
AC C
However, many clinicians have been skeptical of these results because multivariate analyses cannot adjust for the marked clinical differences between VLBW neonates who receive steroids versus those who do not. This problem occurs because VLBW neonates represent a heterogeneous group of infants who range from those of very immature gestational age to those who are more mature but extremely growth retarded.6 In addition, many substantive differences in risk factors for ROP exist in chronically ventilated neonates compared to premature infants
ACCEPTED MANUSCRIPT
without significant pulmonary disease. The purpose of this study was to investigate the association between postnatal steroids and ROP risk by examining ROP risk in a large cohort of preterm neonates born at birth weights
RI PT
at the limit of viability (<500 g). This cohort represents a more homogeneous set of neonates because neonates at these critically low birth weights (and gestational ages) are at the highest vulnerability for a host of neonatal morbidities including ROP and bronchopulmonary dysplasia.
SC
Thus, clinical differences between steroid-treated and untreated neonates can be minimized. We utilized data from Pediatrix BabyStep Clinical Data Warehouse, one of the world’s
M AN U
largest repositories of neonatal data, to examine the incidence of ROP in a large cohort of preterm infants with birthweight of <500 g. This database includes information on >1,120,000 infants who had been cared for by Pediatrix Medical Group providers at US hospitals since 1996.7-8 Data is automatically extracted from the daily medical record notes on each child
TE D
admitted by Pediatrix physicians At the current time, the Pediatrix Medical Group cares for nearly 25% of neonates in more than 300 NICUs across the U.S. The Pediatrix BabySteps Clinical Data Warehouse captures diagnostic and procedure coding data from the Pediatrix
EP
Medical Group electronic health record, which strictly adheres to the American Academy of Pediatrics Perinatal Section Coding Committee Guidelines.7
AC C
Methods
All data entry for the Pediatrix BabySteps Clinical Data Warehouse is verified on an ongoing basis by the Pediatrix Clinical Data Warehouse Information Technology team and by a database manager. The data from the Pediatrix BabySteps Clinical Data Warehouse is annually certified as deidentified and are compliant with the US Health Insurance Portability and Accountability Act of 1996. The data has been approved for use in research studies by the Western Institutional
ACCEPTED MANUSCRIPT
Review Board. In addition, an exemption for any additional institutional review board approval for use of the deidentified dataset in this study was granted by the Michigan State University Institutional Review Board.
RI PT
The following 3 inclusion criteria were used for this study population were: (1) birth weight <500 g; (2) hospitalization survival (ie, discharged from hospital alive); (3) ophthalmic ROP examination results available.
SC
The diagnosis and stage of ROP in the medical records was assigned by the board-
certified ophthalmologists who performed the retinal ROP examinations. The retinal findings
M AN U
were standardized by classifying the examination results according to the International Classification of Retinopathy of Prematurity.9-11 In most cases, multiple ROP examinations had been performed for each infant; the most advanced stage of ROP noted in the medical record was recorded as the infant’s ROP stage. Although infants may have received additional ROP follow-
TE D
up after hospital discharge, no post-discharge examination results were available for this study. Given that the average age at discharge was at a postmenstrual age close to term, however, it is highly probable that ROP severity would have reached its peak for the vast majority of the study
EP
participants before hospital discharge.
Study participants, born between 1996 and 2013, had been NICU inpatients at 167
AC C
different hospitals across the United States. Our study dataset included binary data (yes/no) on steroid exposure during pregnancy (ie, antenatal steroid) and after birth (postnatal steroid exposure). Steroid exposure included exposure to any type of steroids (betamethasone, dexamethasone, hydrocortisone, prednisolone, prednisone) for any indication. Also, the neonate may have been exposed to more than one type of steroid and for differing durations of treatment. Statistical significance for unadjusted comparisons between groups was determined with
ACCEPTED MANUSCRIPT
Pearson’s χ2 test or t test (for parametric comparison of means) or Wilcoxon rank sum test (for nonparametric comparison of median Apgar scores). Multivariate models were adjusted for several covariates. To adjust for secular trends in steroid use over the years, different steroid
RI PT
practices in different hospitals and differences in ROP assessment by different ophthalmologists, year of birth, transfer status (born in NICU hospital or transferred in from another hospital), and NICU hospital facility were included as covariates. Because gestational age, birthweights and
SC
oxygen exposure/ventilation (number of days FiO2 >0.21, days on high-flow nasal cannula, days on ventilation, and days on continuous positive airway pressure [CPAP]) are key risk factors for
M AN U
ROP, these were included as covariates. Comorbidities and other demographics that significantly differed between the steroid treated and untreated groups were included as covariates (patent ductus arteriosus, brain hemorrhage, necrotizing enterocolitis, bronchopulmonary dysplasia, sepsis, Apgar scores, and chronological age at hospital discharge). Antenatal steroid exposure
Results
TE D
and sex were also included as covariates.
Data from 1,472 infant entries from the Pediatrix BabySteps Clinical Data Warehouse repository
EP
satisfied all three inclusion criteria and were used to compile the study dataset. Table 1 compares the characteristics of the postnatal steroid-treated and untreated neonates. Of the 1,472 infants,
AC C
1059 (71.9%) received postnatal corticosteroids and 413 (28%) did not. There were no significant differences between groups with regard to race, sex, or antenatal steroid exposure (ie, steroid exposure during fetal life). There were small but statistically significant differences between groups in mean birthweight, mean gestational age, and median Apgar scores. As expected, there were large, significant differences between groups in variables relating to oxygen exposure.
ACCEPTED MANUSCRIPT
Table 2 provides the unadjusted incidence of ROP and other comorbidities of steroidtreated and nontreated neonates. The overall incidence of ROP (of any stage) for the entire cohort was 76.6% (1128/1472), and the overall incidence of advanced stage ROP (stages 3, 4, or
RI PT
5) was 31.3%. The incidence of any ROP was significantly higher (P < 0.0001) in steroid-treated infants (80.5%) than in nontreated infants (66.8%); the incidence of advanced-stage ROP was also significantly higher (P < 0.0001) in the former (35.3%) compared to the latter group
SC
(21.1%). Steroid-treated infants also had a significantly higher incidence of bronchopulmonary dysplasia, sepsis, patent ductus arteriosus, and intracranial hemorrhage compared to nonsteroid
M AN U
treated neonates.
Table 3 provides the univariate and multivariate effect of postnatal steroid use and sepsis on the odds of ROP incidence for any ROP and for advanced ROP. All multivariate logistic regression included the covariates of postnatal steroids, bronchopulmonary dysplasia, and sepsis
TE D
as well as the covariates of hospital facility, transfer status, antenatal steroid use, birth year, gestational age, birthweight, patent ductus arteriosus, Apgar scores, brain hemorrhage, discharge age, and sex as well as each type of O2 exposure (number of days on FiO2, artificial ventilation,
EP
high flow nasal cannula, and CPAP).
In multivariate analysis there was a 1.6 times greater odds (95% CI, 1.2-2.2) of any ROP
AC C
for steroid-treated infants compared to nontreated infants and 1.7 times greater odds (95% CI, 1.3-2.3) of advanced ROP for steroid-treated infants. In multivariate analysis the odds were 1.5 times greater of any ROP for infants with sepsis (95% CI, 1.1-2.0) and 1.4 times greaterof advanced ROP (95% CI, 1.1-1.8). Discussion The current study indicates that there is a higher risk of ROP in steroid-treated infants than in
ACCEPTED MANUSCRIPT
those infants not treated with steroids. The association between postnatal steroid use in VLBW infants and ROP has been long debated. Some randomized controlled trials involving postnatal steroid show trends toward increased ROP risk but few show any significant difference.1-2 In
RI PT
general, the postnatal steroid randomized controlled trials assess ROP as a secondary outcome and are not primarily designed to detect ROP risk. For example, most randomized controlled trials do not stratify birth weight below 1500 g and there are marked differences between infants
SC
born at different gestational ages in this group overall.1-2 As a result, ROP risk analysis is
within the overall VLBW population.12
M AN U
problematic since birthweight persists as a fundamental, independent ROP risk factor, even
In addition to the lack of birth weight stratification, diagnostic ROP protocols in some of the randomized control trials veer from accepted standards.13 For example, Suske and colleagues14 assessed ROP via a single retinal examination performed at 4 weeks’ postpartum
TE D
age. Because ROP can occur and progress anytime until retinal maturation,13 their method would likely underestimate both ROP incidence and severity in any study population. Another randomized controlled study assessed for “acute ROP” but offered no further definition for this
EP
diagnosis.15
Two separate recent Cochrane reviews performed meta-analyses (integrating data from
AC C
several small randomized controlled trials to examine the effect of the timing of steroid initiation on risk of various outcomes (including ROP) for preterm infants.1-2 These reviews reported that late steroids (ie, initiated after 1 week of life) significantly increased ROP risk, whereas early steroids (ie, initiated after birth but before one of life) significantly decrease risk.1-2 Because the upper limit of their 95% confidence interval limit for early steroid meta-analysis bordered on 1 (0.97 to be exact), 2 the Cochrane meta-analysis for the effect of early initiation of steroids on
ACCEPTED MANUSCRIPT
ROP risk should be interpreted with caution. In addition, the Cochrane meta-analyses incorporated data from some aforementioned randomized controlled studies with substandard ROP methodology that added additional concerns about the validity of their conclusions with
RI PT
respect to ROP risk.14-15 Thus, to date, the available meta-analyses do not provide convincing evidence that postnatal steroids are either harmful or beneficial with respect to ROP.
Given these issues with the postnatal steroid randomized controlled trials and meta-
SC
analyses, the present study undertook a large retrospective review in order to shed light on the ROP-steroid association. We specifically chose our study population to consist of neonates with
M AN U
birthweight of <500 g because this category of neonates is generally more homogeneous than those that include larger birth weight infants; thus, clinical differences between treated/untreated groups, though not entirely eliminated, are minimized. In addition, our study population was limited to infants who survived hospitalization in order to eliminate death as a competing
TE D
outcome with ROP. In other words, because all infants survived, all had an “equal opportunity” for ROP to have either developed or progressed. However, because survival was an inclusion criteria, our study population was female-weighted (male sex is known to be a strong risk factor
EP
for death in extremely premature infants).16 For this reason, sex was included as a covariate in our multivariate analyses.
AC C
Given that bronchopulmonary dysplasia is one of the main indications for corticosteroids,
we found large expected differences between our groups (treated/untreated) with respect to oxygen administration. We adjusted for oxygen exposure in our multivariate analyses. It is important to remember, however, that oxygen exposure per se is not the key risk factor in the development of ROP; rather, tissue oxygen content (paO2 or blood saturation) is the critical element. Although oxygen exposure was used as a marker for hyperoxemia, it is important to
ACCEPTED MANUSCRIPT
note that the two are not always the same. Despite statistically significant differences in birth weight, gestational age, and other characteristics/morbidities between the treated and untreated groups, the magnitude of the
RI PT
differences between groups was generally small; nevertheless, they were adjusted for in
multivariate analysis. Of note, some aspects of the International Classification of Retinopathy of Prematurity (ICROP) were updated by a consensus statement published in 2005.11 Thus, study
SC
participants born before 2005 were staged by the ICROP of 1987 and those born after 2005 were staged according to the updated ICROP. That said, the modifications to the original ICROP were
M AN U
relatively few and did not affect whether or not an infant would be diagnosed with ROP and also did not affect whether or not the ROP would have been classified as early versus late stage. Thus, our results are not affected by this change. Nonetheless, we adjusted for year of birth in multivariate analyses.
TE D
Despite equal medians of Apgar scores at 1 minute in the treated and untreated group, a statistical difference (P < 0.0001) between these groups was detected for this variable by the Wilcoxon rank sum test (ie, lower scores below median for steroid treated group compared to
EP
non-treated group) This finding can occur because the Wilcoxon test ranks all of the observations from both groups and then sums the ranks from one of the groups and compares this to the
test.
AC C
expected rank sum. In other words, the nonparametric Wilcoxon test is not specifically a median
With multivariate analysis, the use of postnatal steroids was a significant ROP risk factor
for critically low birth weight newborns. It yielded between 1.6 and 1.7 times an increase in the odds of ROP. In neonates of birth weight <1500 g, ROP incidence approximates 25% and in neonates with birth weight <1000 g, ROP incidence exceeds 50%.17-18 Given this high incidence
ACCEPTED MANUSCRIPT
of ROP in the VLBW population, even a modest increase in the odds of ROP puts a large number of additional VLBW infants at risk for lifelong visual disability. We identified sepsis to be another significant ROP risk factor for any ROP as well as for
RI PT
advanced ROP. Sepsis has been correlated with ROP in other VLBW studies as well.19This finding is relevant because the aforementioned Cochrane meta-analysis showed a trend toward increased sepsis in steroid-treated neonates1; thus, adverse effects of steroids, such as increase in
SC
sepsis, may additionally raise ROP risk.
The retrospective nature of this study prevents drawing any causal inferences. In addition,
M AN U
the binary nature of the steroid data does not permit for the evaluation of what steroid type, mode of delivery, timing, duration, or dosage are associated with ROP risk. Despite these limitations, this study of a large database of critically low birth weight survivors indicates that steroid-treated infants have a modest but significantly increased risk for ROP. This observation is of potential
TE D
clinical significance, because children with a history of ROP are not only at increased risk for visual impairments from the ROP itself but are also at increased risk for developing other ocular
AC C
EP
disorders later in life.20-23
ACCEPTED MANUSCRIPT
References 1.
Doyle LW, Ehrenkranz RA, Halliday HL. Late (>7 days) postnatal corticosteroids for chronic lung disease in preterm infants. Cochrane Database Syst Rev 2014;5:CD001145. Doyle LW, Ehrenkranz RA, Halliday HL. Early (<8 days) postnatal corticosteroids for
RI PT
2.
preventing chronic lung disease in preterm infants. Cochrane Database Syst Rev 2014;5:CD001146.
Watterberg KL. Policy statement—postnatal corticosteroids to prevent or treat
SC
3.
bronchopulmonary dysplasia. Pediatrics 2010;126:800-808.
Batton DG, Roberts C, Trese M, Maisels MJ. Severe retinopathy of prematurity and
M AN U
4.
steroid exposure. Pediatrics 1992;90:534-6. 5.
Ramanathan R, Siassi B, deLemos RA. Severe retinopathy of prematurity in extremely low birth weight infants after short-term dexamethasone therapy. J Perinatol
6.
TE D
1995;15:178-82; quiz 183-74.
Arnold CC, Kramer MS, Hobbs CA, McLean FH, Usher RH. Very low birth weight: a problematic cohort for epidemiologic studies of very small or immature neonates. Am J
7.
EP
Epidemiol 1991;134:604-13.
Spitzer AR, Ellsbury D, Clark RH. The Pediatrix BabySteps(R) Data Warehouse—a
AC C
unique national resource for improving outcomes for neonates. Indian J Pediatr 2015;82:71-9.
8.
Olsen IE, Groveman SA, Lawson ML, Clark RH, Zemel BS. New intrauterine growth curves based on United States data. Pediatrics 2010;125:e214-24.
9.
Committee for the Classification of Retinopathy of Prematurity. An International Classification of Retinopathy of Prematurity. Arch Ophthalmol 1984;102:1130-34.
ACCEPTED MANUSCRIPT
10.
International Committee for the Classification of the Late Stages of Retinopathy of Prematurity. An International Classification of Retinopathy of Prematurity. II. The classification of retinal detachment. Arch Ophthalmol 1987;105:906-12. The International Classification of Retinopathy of Prematurity revisited. Arch
RI PT
11.
Ophthalmol 2005;123:991-9. 12.
Miller MM, Revenis ME, Lai MM, et al. Risk and clinical course of retinopathy of
Fierson WM. Screening examination of premature infants for retinopathy of prematurity. Pediatrics 2013;131:189-95.
14.
M AN U
13.
SC
prematurity in 78 infants of gestational age 22-25 weeks. J AAPOS 2014;18:266-70.
Suske G, Oestreich K, Varnholt V, Lasch P, Kachel W. Influence of early postnatal dexamethasone therapy on ventilator dependency in surfactant-substituted preterm infants. Acta Paediatr 1996;85:713-18.
Yeh TF, Torre JA, Rastogi A, Anyebuno MA, Pildes RS. Early postnatal dexamethasone
TE D
15.
therapy in premature infants with severe respiratory distress syndrome: a double-blind, controlled study. J Pediatr 1990;117(2 Pt 1):273-82. Tyson JE, Parikh NA, Langer J, Green C, Higgins RD. Intensive care for extreme
EP
16.
prematurity--moving beyond gestational age. N Engl J Med 2008;358:1672-81. Lermann VL, Fortes Filho JB, Procianoy RS. The prevalence of retinopathy of
AC C
17.
prematurity in very low birth weight newborn infants. J Pediatr (Rio J) 2006;82:27-32.
18.
Wheeler DT, Dobson V, Chiang MF, et al. Retinopathy of prematurity in infants weighing less than 500 grams at birth enrolled in the early treatment for retinopathy of prematurity study. Ophthalmology 2011;118:1145-51.
19.
Kavurt S, Ozcan B, Aydemir O, Bas AY, Demirel N. Risk of retinopathy of prematurity
ACCEPTED MANUSCRIPT
in small for gestational age premature infants. Indian Pediatr 2014;51:804-6.
20.
Hartnett ME, Gilbert MM, Hirose T, Richardson TM, Katsumi O. Glaucoma as a cause
RI PT
of poor vision in severe retinopathy of prematurity. Graefes Arch Clin Exp Ophthalmol 1993;231:433-8. 21.
Lee RW, Mayer EJ, Markham RH. The aetiology of paediatric rhegmatogenous retinal
22.
SC
detachment: 15 years experience. Eye (Lond) 2008;22:636-40.
Yang CS, Wang AG, Sung CS, Hsu WM, Lee FL, Lee SM. Long-term visual outcomes
years. Eye (Lond) 2010;24:14-20.
Krolicki TJ, Tasman W. Cataract extraction in adults with retinopathy of prematurity.
EP
TE D
Arch Ophthalmol 1995;113:173-77.
AC C
23.
M AN U
of laser-treated threshold retinopathy of prematurity: a study of refractive status at 7
ACCEPTED MANUSCRIPT
Table 1. Demographics and characteristics of postnatal steroid-treated and untreated neonates in critically low birth weight cohort (birth weight <500 g; N =1472) No steroids N = 413
>0.05a a >0.05
RI PT
134 (32.4) 157 (38.0) 135 (32.7) 12 (2.9) 109 (26.4) 311 (75.8) 4 25.2 ± 2.2 461.8 ± 40.4 14.8 ± 6.5 68.0 ± 53.9 14.2 ± 19.4
P value
a
>0.05 b <0.0001 c <0.0001 c <0.05 <0.001c c <0.0001 c <0.001
SC
Postnatal steroids-treated N = 1059 Sex: male (%) 415 (39.1) White 449 (42.4) Black 345 (32.6) 34 (3.2) Asian Other/Unknown 231 (21.9) +antenatal steroid exposure (%) 831 (78.6) Median 1-minute Apgar score 4 GA, weeks, mean ± SD 24.6 ± 1.5 BW, g, mean ± SD 456.6 ± 43.2 CA at discharge, weeks, mean ± SD 17.6 ± 6.6 No. days FiO2 >.21, mean ± SD 109.8 ± 52.5 No. days on high-flow nasal cannula, mean 18.3 ± 22.2 ± SD No. days on ventilation, mean ± SD 62.5 ± 37.9 No. days on CPAP, mean ± SD 17.0 ± 19.4
c
33.1 ± 33.0 <0.001 c 14.7 ± 19.9 <0.05
M AN U
Characteristic
BW, birth weight; CA, chronological age; CPAP, continuous positive airway pressure; GA, gestational age; SD, standard deviation. a
Chi squared. Wilcoxon rank square test. c t test.
AC C
EP
TE D
b
ACCEPTED MANUSCRIPT
Table 2. Comparison of the prevalence of ROP and other comorbidities in neonates of birth weight <500 g treated with postnatal steroids and not treated with steroids (N = 1,472)
852 (80.5) 374 (35.3)
276 (66.8) 87 (21.1)
<0.0001 <0.0001
122 (11.5) 735 (69.4) 615 (58.1) 95 (9.0) 55 (5.2) 782 (73.8) 402 (37.9)
39 (9.4) 227 (55.0) 185 (44.8) 29 (7.0) 15 (3.6) 203 (49.1) 107 (25.9)
>0.05 <0.0001 <0.0001 >0.05 >0.05 <0.001 <0.0001
M AN U
ROP, retinopathy of prematurity.
SC
ROP, n (%) Any stage Advanced (stage 3, 4, 5) Other disorders, n (%) Congenital heart disorders Patent ductus arteriosus Intracranial hemorrhage (any type) Necrotizing enterocolitis (medical) Necrotizing enterocolitis (surgical) Bronchopulmonary dysplasia Sepsis a 2
AC C
EP
TE D
χ test.
a
RI PT
Postnatal steroid- No postnatal P value treated steroids N = 1,059 N = 413
ACCEPTED MANUSCRIPT
Postnatal steroids Sepsis
Postnatal steroids Sepsis
Odds of ROP, any stage (95% CI) Univariate logistic Multivariate logistic a regression regression 2.0 (1.6-2.6) 1.6 (1.2-2.2) 2.0 (1.5-2.6) 1.5 (1.1-2.0) Odds of advanced ROP (95% CI) Univariate logistic Multivariate logistic a regression regression 2.0 (1.6-2.7) 1.7 (1.3-2.3) 1.8 (1.5-2.3) 1.4 (1.1-1.8)
CI, confidence interval; ROP, retinopathy of prematurity.
SC
a
RI PT
Table 3. Effect of postnatal steroids on odds of ROP incidence in critically low birthweight neonates (birth weight <500 g)
AC C
EP
TE D
M AN U
Multivariate logistic regressions included covariates of postnatal steroids, bronchopulmonary dysplasia, sepsis, hospital facility, transfer status, antenatal steroid use, birth year, gestational age, birthweight, patent ductus arteriosus, Apgar scores and each type of O2 exposure (no. days on FiO2, vent, high-flow nasal cannula, CPAP), brain hemorrhage, discharge age, and sex.