Purpose of study Prolonged mechanical ventilation (MV) of premature infants leads to lung injury. Frequently, the brain is injured. Our studies, using preterm lambs, indicate that MV from 3 d to 21 d shifts balance to more apoptosis and less proliferation of neurons, immature and mature oligodendrocytes, and astrocytes. These results are consistent with less brain-derived neurotrophic factor in the same brain tissue. Methods used We hypothesised that progressively longer period of MV will alter presence of neural stem cells and neuronal progenitor cells, as well as capillaries in the brain. Preterm lambs, treated with antenatal steroids and postnatal surfactant, were managed by MV or high-frequency nasal ventilation (HFNV) for either 3 d or 21 d (n=4/group). We use HFNV as the positive gold-standard for alveolar formation in the lung. At the end of 3 d or 21 d, cortical brain tissue from the temporal lobe was fixed. We used immunohistochemistry to localise Nestin-positive neural stem cells, double cortin (DCX)-positive neuronal progenitor cells, and p-glycoprotein-positive capillaries. We used stereology to quantify surface density of each cell type in cortical grey matter, using systematic, uniform, random sampling. Summary of results Initial results for n=2/group are summarised. Surface density (Sv; cm-1) for Nestin-positive cells appears to increase from 3 d to 21 d of MV (average Sv 5 and 46, respectively), and from 3 d to 21 d HFNV (average Sv 7 and 19, respectively). Sv for DCX-positive cells appears to decrease from 3 d to 21 d of MV (average Sv 94 to 71, respectively), and from 3 d to 21 d HFNV (average Sv 95 and 54, respectively). Sv for capillaries also appears to decrease from 3 d to 21 d of MV (average Sv 84 to 67, respectively), and from 3 d to 21 d HFNV (average Sv 101 and 71, respectively). Conclusions Contrary to our hypothesis, these initial results suggest that preterm birth and any ventilatory support may change the proportion of neural stem cells (Nestin-positive cells), neuronal progenitor cells (DCX-positive cells), and capillaries (p-glycoprotein-positive microvessels) in the brain as the duration of respiratory support increases. Analyses are ongoing to increase sample size. HL110002.

122 Table 1 EGA (wks) BWT (g) Male (%) Fetal Dx Defect >2 Patch BiPDA iNO Dopa NO; n=56 38.2 (1.5) 3137 (481) 33 (59%) 23 (41%) 12 (21%) 6 (11%) 30 (54%) 30 (54%) 34 (61%) MIL; n=31 38.1 (1.5) 3041 (420) 19 (61%) 23 (74%) 12 (39%) 8 (26%) 27 (87%) 27 (87%) 29 (94%) Conclusions MIL use was associated with similar improvement in OI, PAP and cardiac function as NO milrinone therapy. No adverse effects were associated with MIL use. Randomised trials are needed. Neonatology–general II Concurrent session 3:15 PM Thursday, January 25, 2018 123 DOES INDOMETHACIN USED FOR TREATMENT OF PATENT DUCTUS ARTERIOUSUS (PDA) LEAD TO NECROTIZING ENTEROCOLITIS AND/OR INTESTINAL PERFORATION IN PREMATURE INFANTS? B Hwee*, A Wu, A Bracamonte, B Chun, J Chadwick, K Lee, O Shayegh, T Goel, B Afghani. University of California, Irvine, Irvine, CA; University of California, Los Angeles, Westwood, CA; CHOC Hospital of Orange, Orange, CA 10.1136/jim-2017-000663.123 Purpose of study It is unclear whether use of indomethacin (INDO) for treatment of PDA in premature infants has an effect on the development of necrotizing enterocolitis (NEC) or spontaneous intestinal perforation (SIP). The purpose of this study is to investigate if the post-natal administration of INDO is associated with an increased risk of developing NEC and SIP in premature infants. Methods used A literature review using Pubmed and Google Scholar with the keywords Patent Ductus Arteriosus, Indomethacin, Intestinal Perforation and Necrotizing Enterocolitis was conducted. Only studies published after 1990 with premature infants who were administered INDO after birth as treatment for PDA were included. Studies without a control group (no INDO) were excluded. Summary of results Of the 23 articles, only 7 satisfied our inclusion criteria (see table 1). The main reason for exclusion was lack of control group or use of treatments other than INDO. The dosage and timing of INDO administration were variable among studies. Only 1 of 7 studies found a significant increase in NEC occurrence in the INDO group. Most Abstract 122 Figure 1 Change in OI & PAP from baseline: milrinone vs none Abstracts122 Figure 1 Change in OI & PAP from baseline: milrinone vs none Abstracts 116 J Investig Med 2018;66:62–287 studies found a correlation with INDO and development of SIP, especially if INDO was given early or combined with steroids. Conclusions Our review does not suggest a link between INDO and development of NEC; however, there may be a link between SIP in patients who receive INDO. Prospective studies that control for other variables, such as the effects of dosage and timing of INDO as well as factors such as steroids, other medications, and feeding are warranted. 124 OXYGEN CONTENT AND ITS ASSOCIATION WITH RETINOPATHY OF PREMATURITY P Jung, L Geoffrion*, G Truong, M Goldstein. Loma Linda University, Loma Linda, CA 10.1136/jim-2017-000663.124 Purpose of study Hyperoxia is a significant risk factor for development of retinopathy of prematurity (ROP). Oxygen saturation targets have been widely studied with variable results. We evaluated whether oxygen content was associated with development of ROP. Methods used We retrospectively reviewed charts of infants admitted between January 2013 and December 2016 who were eligible for ROP screening (gestational age £30 weeks or birth weight £1500 grams) and had arterial blood gases obtained in the first 2 weeks of life. Arterial oxygen content values were collected. Outcomes measured were incidence and severity of ROP. Dell Statistica was used for all statistical analysis. Summary of results There were 305 total infants and 132 had ROP (43.3%). Prevalence of stage 1 was 19.0%, stage 2 was 15.4%, and stage 3 was 8.9%; there were no higher stages of ROP in this population. Decreasing oxygen content within the first 2 weeks of life was significantly associated with higher severity of ROP. Infants without ROP had an average oxygen content of 16.0±6.0 (standard deviation). Average oxygen content for stage 1 was 14.8±4.6, stage 2 was 14.6±4.7, and stage 3 was 13.6±2.4 (p=0.004). Conclusions In our population, higher oxygen content levels within the first 2 weeks of life was associated with decreasing severity of ROP development. As previous studies have shown worse ROP with higher saturation targets, this finding is unexpected. Based on our findings, though, oxygen saturations may only be a component of the mechanism leading to the development ROP. There may be further factors not yet studied that may contribute to its development. Further studies are required. Abstract 123 Table 1 First author and year Gestational age and birth weight (Grams) Timing of INDO administration Subject INDO w/NEC Control group w/NEC Subject INDO w/IP Control group w/IP NEC P-value IP P-value Grosfeld 1996 3 to 7 days of age 90/252 (35%) 105/764 (13.7%) 27/252 (10.7%) 13/764 (1.7%) p<0.02 p<0.05 Fujii 2002 ETG=Early Treatment: First 2 days of life ETG=10/30 (33%) STG=9/32 (28%) ETG=6/30 (20%) STG=0/32 (0%) p=0.659 p=0.011 STG=Standard Treatment: After 2 days of life O’Donovan 2003 As needed for significant PDA 14/108 (13.0% I) 7/50 (14.0% L) 5/108 (4.6% I) 6/50 (12% L) NS NS 8/66 (12.0% I-L) 4/66 (6.1% I-L) Dolberg 2005 24 to 34 weeks NR 109/1183 (9.2% PDA-I) 192/4488 (4.3% No PDA) 37/373 (9.9% PDA only) NR NR Presence of PDA itself (not I) was significant for NEC NR McPherson 2008 As needed for significant PDA 14/228 (6.1%) 47/628 (7.5%) 14/228 (6.1%) 4/628 (0.6%) NS p Pumberger 2014 First 24 hours of