King's College London

Research portal

Prognostic value of the oxygenation index to predict survival and timing of surgery in infants with congenital diaphragmatic hernia

Research output: Contribution to journalArticle

Yew Wei Tan, Kamal Ali, Gwendolyn Andradi, Lekshmi Sasidharan, Anne Greenough, M Davenport

Original languageEnglish
JournalJournal of Pediatric Surgery
Early online date27 Dec 2018
Publication statusE-pub ahead of print - 27 Dec 2018

King's Authors


Aims: To compare the mean oxygenation index on day1 (MOId1) with the best OI on day1 (BOId1) in predicting 30-day mortality in congenital diaphragmatic hernia (CDH). In addition, to determine whether serial OIs in the first 48 h after birth and preoperative OI (PreOp-OI), were associated with optimal timing of surgery, ventilation requirement and hospital stay in infants with CDH. Methods: The medical records of infants with antenatally diagnosed CDH during 2009–2015 were retrospectively reviewed. Receiver operator characteristic curves were constructed to identify MOId1 and BOId1 cutoff levels to predict 30-day mortality. In those who underwent surgery, the mean OI on each of the first two days (MOId1, MOId2) and PreOp-OI were correlated using Spearman (rs) with the age at surgery, the duration of ventilation and length of stay. Statistical significance was P < 0.05* and < 0.001**. Results: Survivors (n = 44) and nonsurvivors (n = 24) were comparable in gestational age, birth weight and defect laterality. Nonsurvivors had higher median BOId1 (15.4 vs 2.9; P < 0.01) and MOId1 (48 vs 7.5; p < 0.01) than survivors. Mortality was best predicted by two cutoffs [BOId1 > 6 (sensitivity 92%, specificity 89%); and MOId1 > 17 (sensitivity 96%, specificity 96%)]. Forty-four infants underwent surgery at a median postnatal age of five (range 2–19) days. MOId1 and MOId2 both correlated significantly with the age at surgery (rs = 0.4** rs 0.5**) but not ventilation period and length of stay. PreOp-OI correlated significantly with age at surgery, duration of ventilation and length of stay (rs = 0.32* rs = 0.47** rs = 0.37*). A PreOp-OI < 3 was predictive of optimal timing for surgery with improved duration of ventilation (8 vs 22, P = 0.001) and length of stay (26 vs 47, P = 0.004). However, 11/44 (25%) patients would not achieve a PreOp-OI < 3 by day 7 + of life and might still require surgery, one of them died. Conclusion: Both MOId1 and BOId1 are highly predictive of mortality in CDH. Oxygenation indices in the first 48 h poorly predicted the timing of surgery. PreOp-OI < 3 may be a cutoff for optimal timing for surgery in infants for CDH repair. Type of study: Prognostic Study. Level of evidence: Level III.

View graph of relations

© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454