Hyperbilirubinemia and Retinopathy of Prematurity in Preterm Infants: a Retrospective Study.

Retinopathy of prematurity (ROP) is a disease related to prematurity with a multifactorial and only partially known etiopathogenesis.

A better understanding of its risk factors and potential protective factors may lead to the development of better preventive and therapeutic strategies in the future and, thus, ultimately, reduce the incidence of this pathology.

The starting hypothesis of our study is based on the proven physiological antioxidant effect of bilirubin, demonstrated by various in vitro and in vivo studies: it was hypothesized that bilirubin, by its antioxidant action, could exercise a protective role against oxidative stress pathologies typical of prematurity, including ROP.

However, the antioxidant action of bilirubin in humans, particularly newborns, is still controversial. Some studies have demonstrated a role protective of bilirubin against oxidative damage in preterm infants, while others have conversely observed that hyperbilirubinemia is a risk factor for ROP.

It has been hypothesized that bilirubin exerts its antioxidant effect only under conditions of increased oxidative stress, i.e., when the remaining antioxidant mechanisms are no longer sufficient to cope with the production of highly reactive molecules, for example, under conditions of hypoxia and hyperoxia.

According to some studies, there is a threshold within which bilirubin can exert a cytoprotective action, beyond which the harmful effects prevail, especially in the brain. Therefore, it is necessary to have a cohort of large numbers of subjects to define better the role of hyperbilirubinemia in developing ROP, especially in the group of subjects most at risk, such as the preterm.

Among the variables analyzed, the investigators also included the value of fetal hemoglobin expressed as a percentage of total hemoglobin, assuming that lower fetal hemoglobin levels may be associated with a higher incidence of ROP. Fetal hemoglobin represents about 80% of total hemoglobin at birth and subsequently decreases physiologically, replaced by hemoglobin A until it reaches 5% at six months of age.

Transfusions of concentrated red blood cells derived from adults are mainly composed of hemoglobin A, leading to a greater and more rapid reduction in fetal hemoglobin values in newborns. Compared to hemoglobin A, fetal hemoglobin has a greater affinity for oxygen; therefore, lower fetal hemoglobin levels could lead to greater oxygen availability in the tissues and the retina, contributing to hyperoxia, which is known is the basis of the pathogenesis of ROP. With these premises, some studies conducted on small populations have analyzed the association between fetal hemoglobin levels and the onset of ROP with discordant results.

Primary endpoints:

• Number of subjects who developed ROP (divided by severity)
• Number of subjects who developed jaundice
• Duration of jaundice (days) for each subject
• Potential confounding variables: EG, PN, oxygen administration, number of RBC transfusions, sepsis, intraventricular hemorrhage, and necrotizing enterocolitis.

Study design:

Monocentric, retrospective cohort observation.

Inclusion criteria:

• Gestational age < 32 weeks and/or birth weight < 1500g - born or hospitalized at the Neonatology and Neonatal Intensive Care Unit of the Foundation in the period under study (between 01/01/2010 and 12/31/2020).

Exclusion criteria:

Infants who do not meet the study inclusion requirements and who meet at least one of the following criteria will be excluded: - lack of data regarding the development of ROP - lack of data on the development of jaundice.

Procedures of the study:

Data from hospitalized infants in the period between 01 /01/2010 and 31/12/2020 will be collected retrospectively from the database of computerized medical records of the UOC Neonatology and Neonatal Intensive Care of the Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico.

The variables that will be collected are the following:

• general data of the newborn: date of birth, gender, PN, EG, mode of delivery, twins, and any treatments for complications of twins - pregnancy history: the presence of pathologies, complications, and therapies - perinatal data: assessment of adaptation to extrauterine life, need for resuscitation and/or drugs in the delivery room - neonatal parameters and comorbidities: fetal hemoglobin values, intraventricular hemorrhage, respiratory distress syndrome, pulmonary broncho dysplasia, necrotizing enterocolitis, patency of the ductus arteriosus, sepsis, malformations, presence, degree and possible treatment of ROP, presence, duration, and treatment of jaundice - treatments: need, type and duration of ventilation, administration of oxygen, number of transfusions of packed red blood cells The data will be collected from the database of computerized medical records (Neocare) of the UOC of Neonatology and Neonatal Intensive Care Unit of the IRCCS Ca' Granda Foundation Ospedale Maggiore Policlinico.

The source documents include all records of observations or annotations of clinical activities and all reports and records necessary for the evaluation and reconstruction of the clinical history.

The object of the study:

Correlation between neonatal indirect bilirubin jaundice and ROP. Control group: infants who, despite presenting ROP, did not develop jaundice.

Number of patients:

Number of patients planned for the study: 1586.

Statistic Analysis:

Population demographic characteristics and jaundice severity will be presented for the whole population using descriptive statistics.

Continuous variables will be shown using mean (standard deviation) for symmetric variables or median (interquartile range) for skewed variables and number (%) for categorical variables.

Comparing the patients with and without jaundice, categorical variables will be compared using Fisher's exact test. In contrast, continuous variables will be compared using a t-test (symmetric variables) or Mann-Whitney U test (asymmetric variables). Regarding comparing groups with different durations of jaundice, categorical variables will be compared using Fisher's exact test. In contrast, continuous variables will be compared using one-way ANOVA (symmetric variables) or the Kruskal-Wallis test (asymmetric variables).

All tests performed are two-tailed, and p-values < 0.05 are considered significant. All analyses will be performed using R software version 4.0.1 or higher (R Foundation for Statistical Computing, Vienna, Austria).

Financing: None

GCP statement: This study will be conducted in accordance with the protocol, rules of the ICH/GCP (International Council for Harmonization/Good Clinical Practice), and applicable regulations, including the June 1964 Declaration of Helsinki, as amended by the last World Medical Association General Assembly in Seoul in 2008.