Introduction 4 5 2 3 1 In this study, we retrospectively analysed the incidence, severity and causalities of PM in neonates and children >4 weeks of life admitted to our intensive care unit, and we investigated the possible differences between the groups. Material and methods We retrospectively reviewed all records of children diagnosed with PM who were hospitalised in the interdisciplinary neonatal and paediatric intensive care unit of the University Children’s Hospital in Zürich, Switzerland, between January 2000 and September 2006. The patients were divided into two groups according to their age: neonates (under 4 weeks of age) and children (over 4 weeks of age). We were interested in the causes of PM as documented by the treating physicians, the types and results of radiologic investigations performed, any invasive interventions used to treat PM, the severity of the PM and the length of stay in the intensive care unit. Results About 1,200 children are admitted to our intensive care unit per year. The incidence of PM in our intensive care unit was 0.08% for children >4 weeks of age and 0.1% for neonates. In all patients, PM was diagnosed by chest X-ray and all had a positive outcome related to the PM. All five patients with pneumopericardium (PP) did not suffer from any complications (e.g. pericardial tamponade). 1 1 Table 1 Results for the group of children >4 weeks of life Air leak Aetiology Age (years) Intubation (after diagnosis) Pleural drainage Diagnostic tests Days in ICU PM, PP, SE Spontaneous/3 days earlier intensive sport 15.8 No No Chest X-ray 3× 2 PM, PT Traumatic 7.5 3 days Yes CT/chest X-ray 3× 5 PM, SE  Traumatic (lesion in hypopharynx) 1.3 No No CT/chest X-ray 3×, oesophagogram with contrast medium, laryngotracheoscopy 4 PM, PP Iatrogenic: equipment failure with barotrauma during mechanical ventilation 2.3 No No Chest X-ray 2× 2 PM Spontaneous 15.4 No No Chest X-ray/CT 1 PM  Obstructive bronchitis 5.6 No No Chest X-ray 2×/CT 2 PM, PT Obstructive bronchitis 1.9 7 days Yes Chest X-ray 7× 7 Fig. 1 Pneumomediastinum (PM), subcutaneous emphysema (SE) and pneumopericardium (PP) in a 2-year-old intubated patient  The group of children older than 4 weeks stayed in the intensive care unit for a mean of 3.2 days (range 1–7 days), depending on the severity of the PM and the underlying disease. Compared to the neonatal group, the length of stay in the intensive care unit was shorter. However, most neonates stayed in the intensive care unit longer, primarily because of comorbid conditions and not because of the PM. 2 2 Table 2 Results for the group of neonates Air leak Aetiology Birth weight Gestational age (weeks) Mode of delivery Mechanical ventilation before diagnosis Duration of ventilatory support after diagnosis Pleural drainage Days in ICU PM Spontaneous 4,140 g 40 0/7 Vaginal No No No 4 PM, PP, SE Premature lungs, barotrauma 2,150 g 34 4/7 Caesarian section  2 3 days (intubation) No 4 PM, PT, PP Premature lungs, spontaneous or CPAP 2,480 g 35 6/7 Vaginal CPAP 4 days (intubation) Yes 6 PM, PT Spontaneous 3,485 g 38 1/7 Caesarian section No No No 3 PM, PT Spontaneous 3,440 g 37 5/7 Vaginal No 6 hours (CPAP) No 3 PM, PP Spontaneous 2,830 g 39 1/7 Caesarian section No No No 13 PM, PT Spontaneous 3,970 g 38 5/7 Caesarian section No No No 9 PM, PT Pulmonary infection due to maternal infection 3,440 g 38 5/7 Vaginal CPAP 1 day (CPAP) No 5 PM Convulsions or spontaneous 4,130 g 40 5/7 Vaginal No No No 4 Discussion All children with PM had a good outcome without any complications due to air trapping. In the group of children older than 4 weeks, only two children developed a respiratory insufficiency, leading to mechanical ventilation. In both of them, respiratory failure was related to their underlying condition (polytrauma with haematothorax and severe obstructive bronchitis, respectively). These were also the children who stayed longest in the intensive care unit. All other children were treated with oxygen only and stayed in the intensive care unit until they improved clinically and radiographically. Regarding radiologic diagnostics, four patients of the group of children >4 weeks of life had CT scans (three of them had been done in outside clinics from where the patients had been admitted to our intensive care unit). Retrospectively, the utility of the CT scans was put into question, as these scans did not change the patient management. The only patient in whom the CT scan changed management was the child with polytrauma. In this patient, other intrathoracic injuries needed to be ruled out. In the group of neonates, it was much more difficult to find the aetiology of the PM, since all neonates had presented with respiratory disease, and radiologic investigations were partially performed only after the use of CPAP or tracheal intubation. Five of the nine neonates had a spontaneous PM without risk factors, such as mechanical respiratory support (bag mask ventilation after birth, CPAP, mechanical ventilation) or restrictive lung disease. Three of these five babies were delivered by caesarean section. In the remaining four newborns, possible mechanical incidents leading to the air leak could be revealed: mechanical ventilation with high inspiratory pressure, CPAP, pulmonary infection and convulsion. Further investigations are needed to find the aetiology of spontaneous PM in healthy, term neonates. In conclusion, PM in children and neonates has a good prognosis. Mostly, it is associated with extrapulmonary air at other sites. It is diagnosed by chest X-ray alone. Whereas in older children mechanical events leading to the airway rupture can be revealed in most cases, about half of the neonates in our series suffered from PM without obvious reason.