OBJECTIVES:

To evaluate whether the preoperative chest radiographic thoracic area in newborn infants with congenital diaphragmatic hernia was related to the length of mechanical ventilation and the total length of stay and whether chest radiographic thoracic area predicted survival to discharge from neonatal care.

DESIGN:

Retrospective observational cohort study.

SETTING:

Tertiary neonatal unit at King’s College Hospital National Health Service Foundation Trust, London, United Kingdom.

PATIENTS:

Newborn infants admitted with congenital diaphragmatic hernia at King’s College Hospital in a 10-year period (2007-2017).

INTERVENTIONS:

The chest radiographic thoracic area was assessed by free hand tracing of the perimeter of the thoracic area as outlined by the diaphragm and the rib cage and excluded the mediastinal structures and abdominal contents in the thorax and calculated using the Sectra PACS software (Sectra AB, Linköping, Sweden).

MEASUREMENTS AND MAIN RESULTS:

Eighty-four infants with congenital diaphragmatic hernia (70 left-sided) were included with a median (interquartile range) gestation of 36 weeks (34-39 wk). Fifty-four (64%) survived to discharge from neonatal care. In the infants who survived the chest radiographic thoracic area was not related to the length of mechanical ventilation (r = 0.136; p = 0.328) or the total duration of stay (r = 0.095; p = 0.495). The median (interquartile range) chest radiographic thoracic area was higher in infants who survived (1,780 mm [1,446-2,148 mm]) compared with in the deceased infants (1,000 mm [663-1,449 mm]) after correcting for confounders (adjusted p = 0.01). Using receiver operator characteristics analysis, the chest radiographic thoracic area predicted survival to discharge from neonatal care with an area under the curve of 0.826. A chest radiographic thoracic area higher than 1,299 mm predicted survival to discharge with 85% sensitivity and 73% specificity.

CONCLUSIONS:

The chest radiograph in infants with severe congenital diaphragmatic hernia can predict survival from neonatal care with high sensitivity and moderate specificity.