Clinical and Growth Correlates of Retinopathy of Prematurity in Preterm infants with Surgical Necrotizing Enterocolitis and intestinal Perforation

Background: we sought to determine the clinical and growth parameters associated with retinopathy of prematurity (ROP) in infants with necrotizing enterocolitis (NEC) and spontaneous ileal perforation (SIP). Methods: Retrospective cohort study comparing clinical information before and following NEC/SIP onset in neonates with and without severe ROP (Type 1 and 2). Results: Those with severe ROP (32/109, 39.5%) had lower GA, BW, chorioamnionitis, later median onset of ROP diagnosis and received Penrose drain and had higher AKI, poor weight z scores, poor linear growth, longer duration of ventilation and higher FIo2 than those without ROP following NEC/SIP. The GA and diagnosis at later age remained significant for any ROP on multi regression modelling. Conclusion: The surgical NEC/SIP infants with severe ROP were more likely to be younger, smaller, had AKI, had higher oxygen exposure and poor weight gain and linear growth than those without severe ROP.


Introduction
Necrotizing enterocolitis (NEC) is the most common acute gastrointestinal illness, affecting about 5-10% of preterm neonates with a birth weight ≤ 1500 grams [1,2]. NEC remains a leading cause of morbidity due to NEC-associated severe systemic in ammatory response causing multi-organ dysfunction and mortality among preterm neonates and leads to increased health care burden.
Necrotizing enterocolitis is associated with necrosis, in ammation, hemorrhage and reparative changes on intestinal histopathological examination [3]. The hemorrhagic necrosis seen in infants with NEC is most likely due to anormal vasculature and neo angiogenesis in the intestine [4,5]. The Retinopathy of prematurity is also associated with abnormal vascularization because of insulin like growth factor 1 (IGF-1) and vascular endothelial growth factor (VEGF) effect on the retinal angiogenesis [6]. In the Early Treatment for Retinopathy of Prematurity study, ROP developed in 68% of preterm infants < 1251 g and severe retinopathy of prematurity developed in almost 37% of the cases [7].A recent multicenter study has reported 12.8% of preterm infants born less than 28 weeks had severe ROP [8] and 2.5% of infants had bilateral blindness on the long-term follow-up. The surgical NEC and its timing of onset is a signi cant risk factor for the ROP development in preterm infants as shown in published reports [9,10]. Surgical NEC is associated with dysbiosis and poor growth outcomes [11,12]. Recent reports have reported the association between the altered gut microbiome and the retinopathy of prematurity [13]. Poor postnatal weight gain is reported to be a signi cant risk for severe ROP as well [14,15].
There is no comprehensive study evaluating the surgical NEC/SIP characteristics and the growth data in detail in preterm infants with severe ROP before and following the NEC onset. In this single-center, retrospective cohort study, we sought to determine the clinical risk factors and growth characteristics that were associated with ROP before and after the surgical NEC onset in preterm infants.

Population and Study Design
The study was conducted at the Level IV Neonatal unit at University of Mississippi Medical Center after IRB approval with approximately 1000 admissions/year including referrals from throughout the state. All neonates admitted between January 1, 2013 and June 2018, with a diagnosis of NEC (Bell stage III) were included in the study [16]. Neonates diagnosed with medical NEC were excluded from the analysis. The infants included in the study are summarized in Fig. 1.

Demographic data
We collected demographic data including gestational age (GA), birth weight (BW), sex, appropriate for gestational age (AGA) status, race, out born status, mode of delivery, and Apgar scores ≤ 6 at 5 min. We also collected maternal variables including maternal pregnancy-induced hypertension (PIH), chorioamnionitis, and antenatal steroids.
Additional clinical information included mechanical ventilation exposure, presence of patent ductus arteriosus (PDA) and indomethacin/ibuprofen therapy for PDA treatment (before NEC), inotrope (dopamine) use 24 hours after NEC onset. In addition, we collected information on duration, Fio2 requirement, and mode of ventilation (invasive/noninvasive) before and following NEC. We also collected information on the blood culture-proven sepsis at the time of NEC onset, length of stay and mortality. The length of stay was de ned as the total duration of hospitalization from the day of admission until discharge or death due to any cause before hospital discharge.

NEC data
We recorded information on the age (in days) at the time of NEC diagnosis. The diagnosis of NEC was made based on characteristic radiographic ndings including pneumatosis, portal venous gas, and pneumoperitoneum on abdominal X-ray. The frequency of medical and surgical NEC (Bell stage II and III) were also collected [16]. Neonates who died within 48 hours after NEC onset and massive bowel necrosis was found during laparotomy or autopsy were classi ed as having fulminant NEC. At our center, preterm infants with pneumoperitoneum who weigh less than 1 kg at NEC/SIP diagnosis and are hemodynamically unstable are treated rst with a Penrose drain at the bedside but may later receive laparotomy. The timing of laparotomy after placement of Penrose drain was based on clinical deterioration.
Post-operative information such as post-operative ileus days (de ned as infants being NPO after bowel surgery), time to reach full feeds (≥ 120 ml/kg/day) and total parenteral nutrition days were also gathered. The surgical morbidity was classi ed as strictures, stulas, wound dehiscence, surgical site infections (including abscesses), adhesions, and perforations.
Retinopathy of Prematurity Data: We performed an analysis of 109 infants born during 2013-2018 that were admitted to the University of Mississippi Medical Center Neonatal Intensive Care Unit. ROP testing was indicated if the infant was born before 31 weeks gestational age or after 31 weeks if considered high risk. ROP was grouped into three categories: Type 1 ROP, type 2 ROP, and other ROP [10,18]. Type 1 and type 2 ROP are the most severe types and usually require treatment. Any infant with plus disease was categorized as having type 1 ROP. Plus disease indicates dilated veins and tortuous arteries in the posterior pole of the eye. Type 2 ROP is any infant having stage 3 disease. All infants with type 1 ROP were treated with Laser photocoagulation or Avastin. Laser photocoagulation is an ablative treatment that targets avascular regions in order to decrease angiogenic factors such as Vascular Endothelial Growth factor (VEGF) and slow the growth of new abnormal blood vessels. Avastin (Bevacizumab) is a recombinant humanized monoclonal antibody that also targets VEGF and stops neovascularization.

Kidney Function Data
We collected all serum creatinine measurements and daily urine output data starting the day before NEC diagnosis, at NEC onset, and up to 1 week after NEC diagnosis.
Bronchopulmonary dysplasia (BPD) data BPD at 36 weeks corrected gestational age was classi ed as mild, moderate, and severe based on the oxygen requirement at assessment [24]. We collected data on the type of steroid (hydrocortisone/dexamethasone) used during the clinical course.
Growth Outcome data Time intervals include prior to developing NEC, during NEC treatment, post-NEC until anastomosis, after anastomosis, at 36 weeks chronological age, and at discharge. Anthropometric variables include weight, height, weight-for-length, head circumference, and respective z-scores. Sex-speci c Fenton growth charts were used for infants less than 50 weeks old, and gender-speci c WHO corrected for gestational age growth charts were used for infants greater than 50 weeks old.
Brain injury data MRI brain scans are routinely obtained at term equivalent age or before discharge home for all preterm infants weighing less than 1500 grams at birth. The MRI images were scored independently by two pediatric neuroradiologists We used a scoring system of eight scales for white and gray matter injury developed by Woodward et al. [25].The categories of white-matter abnormality were none (a score of 5 to 6), mild (a score of 7 to 9), moderate (a score of 10 to 12), and severe (a score of 13 to 15).

Statistical Methods
In our study, we had analyzed the combined cohort of NEC/SIP and NEC alone. We analyzed all the continuous variables using the Mann-Whitney U-test and summarized with median and inter-quartile range (Quartile 1; Quartile 3). The categorical variables were tested using the Chi-squared test (or Fisher's exact test when cell counts were below 5). The signi cant variables from the bivariate analyses were included in the multiple logistic regression. Adjusted odds ratios were reported as effect size along with 95% con dence interval and P value. Evaluations for signi cant multicollinearity led to birth weight and the corrected gestational age of ROP diagnosis being eliminated from the multivariable modeling process. P values less than 0.05 were considered as signi cant. Statistical analyses were performed in R Statistical Software (version 4.2.1; The R Foundation for Statistical Computing).

Results
The demographic and clinical information of control (n = 54) and the infants with NEC/SIP (n = 109) is summarized in Table 1. Those with surgical NEC had signi cantly lower gestational age, lower birthweight, were less exposed to antenatal steroids (71% vs. 91.8%; p = 0.08), had onset of severe ROP at  ductus arteriosus more often (75% vs. 55%; P = 0.048) and received indomethacin more frequently (22% vs.6.2%, p = 0.045) than those without any ROP. The data is shown in Table 2. There were no signi cant differences in the intestinal histopathology, postoperative features such as time to reach feeds and parenteral nutrition dependance, BPD, white matter and grey matter injury on brain MRI, length of stay and mortality in 2 groups. The data is shown in Table 3. Growth outcomes and Severe ROP The preterm infants with severe (type 1 and type 2) ROP had signi cantly lower length and head circumference z scores before and following NEC. However, weight for length Z scores were signi cantly lower for infants with severe ROP compared to the other group. The data has been summarized in Fig. 2 and Supplemental Table 2.

Regression modelling
On multivariate regression modelling, gestational age (odds ratio 0.41 (95% CI 0.19-0.65); p = 0.004) and day of diagnosis of severe ROP (OR 1.09, 95% CI 1.05-1.17); p = 0.007) following NEC were most likely associated with ROP. The AKI, Fio2 requirement at 7 days of life and the duration of invasive ventilation following NEC were not signi cant. The data has been summarized in Table 5 Table 4.

Growth Outcomes
In the NEC cohort, the weight z scores and weight for length percentiles were signi cantly lower at 36 weeks corrected gestational age for the preterm infants with severe ROP. The length z scores were signi cantly lower before and 4 weeks following NEC onset and before the anastomosis in preterm. The data has been summarized in Fig. 2 and Supplemental Table 2.

Discussion
In our cohort, 55% of infants with surgical NEC/SIP had any ROP and one third of infants had severe ROP.
Type 1 ROP was more common than the Type 2 ROP. Almost one third of infants with severe ROP received laser treatment and one fth of infants received avastin treatment. Only 10% of cases received both laser and the avastin treatment. Those with severe ROP were smaller, younger and were exposed to prenatal risk factors such as PIH, chorioamnionitis and postnatal risk factors including PDA, indomethacin, AKI, more Fio2, invasive and non-invasive mechanical ventilation more frequently. In addition, NEC infants with severe ROP had lower weight z scores and linear growth before and following the NEC onset.
The published reports have shown that the prematurity and the degree and duration of oxygen exposure in uence the incidence of ROP in the preterm infants. In our NEC cohort, infants, infants with ROP were exposed to higher Fio2 and were ventilated invasively and non-invasively before and following the NEC onset in the univariate analysis. However the oxygen exposure was not a signi cant risk on the multivariate analysis, which may be explained due to: 1) other clinical factors play a key role than total oxygen exposure, 2) we failed to model the in -vivo oxygen saturation accurately 3) small sample size in the regression models as reported by Chen et al [26]. The studies have demonstrated the relationship between the oxygen and the ROP with phase I (hyperoxia-induced vasoconstriction and ischemic injury) and phase II (vascular endothelial growth factor-driven Vaso proliferation) of the disease [6].
In our cohort, the infants with severe ROP were younger (23.8 weeks vs. 27.3 weeks) and had lower birth weight (625 grams vs.935 grams) than those without severe ROP as reported in previous published report [27]. Studies done in discordant twin pairs have reported that gestational age is a better predictor of ROP severity than birth weight [28]. In a prospective study done in Australian and New Zealand Neonatal Network Darlow et al had reported that prematurity was the dominant risk factor, with infants with GA of <25 weeks having 20 times greater odds of severe ROP than infants with GA of 28 weeks. Birth weight for GA also had a "dose-response" effect; the more growth-restricted infants had greater risk, with infants below the 3rd percentile of weight for GA having 4 times greater odds of severe ROP than those between the 25th and 75th percentiles. Male gender was also a signi cant risk factor (odds ratio: 1.73; 95% con dence interval: 1.25-2.40) [29]. We did not see any sex difference in our study.
In our study , infants with severe ROP had poor linear growth at evidenced by lower length and head circumference z scores before and following NEC and were more exposed to chorioamnionitis which is similar to a recent prospective study reporting that slower length gain and maternal chorioamnionitis was associated with delayed regression and complete vascularization of retina in preterm infants [30].
In our NEC cohort, the weight z scores and weight for length percentiles were signi cantly lower at 36 weeks corrected gestational age for the preterm infants with severe ROP. As reported in published reports, the poor weight gain postnatally has been associated with severe ROP [31][32][33]. Postnatal weight gain is a surrogate indicator of insulin-like growth factor 1 (IGF-1), and a persistent lower serum IGF-1 in preterm infants is associated with poor weight gain [34]. Low serum IGF-1 also causes insu cient activation of retinal VEGF resulting in poor retinal vascular growth and the development of ROP [35].
In our cohort, AKI by creatinine following NEC onset was most likely associated with the severe ROP on bivariate analysis in infants with NEC/SIP which may most likely be explained due to systemic in ammation initiated secondary to NEC affecting kidneys and retina leading to multiple systemic morbidities. Several studies have shown that preterm infants with surgical NEC have severe white matter injury on the brain MRI, higher serum pro-in ammatory markers, and poor neurodevelopmental outcome at two years of corrected age [25,[36][37][38][39]. Animal studies have reported surgical NEC leads to systemic in ammation and causes neuronal injury via microglial activation, in ammatory pathway activation, and brain barrier disruption [40][41][42][43]. Mechanistically, we hypothesize that severe acute kidney injury in neonates with surgical NEC may exacerbate the injury by acting as a catalyst or modi er of retinal in ammation. Further studies are needed to understand the role of severe kidney injury with ROP in neonates with NEC.
Our study's strengths include a comprehensive evaluation of clinical and growth factors most likely associated with severe ROP. Our study has important limitations. First, this was a single-center experience, reducing the study's generalizability. Also in our cohort, most neonates with NEC were African American. While this is partly due to race distribution in Mississippi, this may also be related to adverse social determinants of health and genetic risk for NEC. Second, sample size limits our power to detect associations between clinical factors, NEC, and ROP. Finally, the small sample size coupled with multiple factors, outcomes, and comparisons may result in type I errors.
In conclusion, data shows that any ROP was diagnosed in 55% cases of infants with NEC/SIP and Type 1 ROP (38%) was more common and received both laser/ avastin therapy in 10% of cases, followed by type 2 in preterm infants with surgical NEC. Those with severe ROP were smaller, younger and were exposed to prenatal risk factors such as PIH, chorioamnionitis and postnatal risk factors including PDA, indomethacin, AKI, more Fio2, invasive and non-invasive mechanical ventilation more frequently. In addition, NEC infants with severe ROP had lower weight z scores and linear growth before and following the NEC onset. There is need to develop and implement strategies including aforementioned clinical risk factors to identify the NEC infants at greater risk of severe ROP to have better short term and the longterm eye outcomes. Weight gain, linear growth and nutrition should be closely monitored in preterm infants with surgical NEC/SIP at higher risk of severe ROP to improve the ophthalmic outcomes.

Declarations
Author Contribution: PMG designed the study; PMG, RR, ACM, JW, DZ, AS, DS, VW, MAYA collected, analyzed the data and wrote the manuscript. All the authors contributed to and approved the manuscript.
Con icts of interest: The authors disclose no con icts.  Table   Table 5 is not available with this version Figures Figure 1 Legend not included with this version.

Figure 2
Legend not included with this version.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.