Article Text

Download PDFPDF

Tension pneumocephalus induced by high-flow nasal cannula ventilation in a neonate
  1. Alicia Iglesias-Deus1,
  2. Alejandro Pérez-Muñuzuri1,
  3. Olalla López-Suárez1,
  4. Pilar Crespo2,
  5. Maria-Luz Couce1
  1. 1Neonatal Unit, Department of Pediatrics, Hospital Clínico Universitario de Santiago, IDIS (Health Research Institute of Santiago de Compostela), Santiago de Compostela, Spain
  2. 2Neonatal Unit, Department of Pediatrics, Complejo Hospitalario Universitario de Pontevedra, Pontevedra, Spain
  1. Correspondence to Dr Alicia Iglesias Deus, Neonatal Unit, Department of Pediatrics, Santiago de Compostela University Hospital, Travesía Choupana, s/n, Santiago de Compostela 15706, Spain; Alicia.Iglesias.Deus{at}


The use of high-flow nasal cannula (HFNC) therapy as respiratory support for preterm infants has increased rapidly worldwide. The evidence available for the use of HFNC is as an alternative to nasal continuous positive airway pressure (CPAP) and in particular to prevent postextubation failure. We report a case of tension pneumocephalus in a preterm infant as a complication during HFNC ventilation. Significant neurological impairment was detected and support was eventually withdrawn. Few cases of pneumocephalus as a complication of positive airway pressure have been reported in the neonatal period, and they all have been related to CPAP. This report reinforces the need to be aware of this rare but possible complication during HFNC therapy, as timely diagnosis and treatment can prevent neurological sequelae. We also stress the importance of paying close attention to flow rate, nasal cannula size and insertion, and mouth position, and of regularly checking insertion depth.

  • Cranial cavity
  • Air
  • Positive pressure ventilation
  • Non-invasive ventilation
  • Newborn

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

What this study adds?

  • Pneumocephalus is a complication of high-flow nasal cannula ventilation.


Pneumocephalus is the presence of air within the cranial cavity. Its main causes (75–90% of cases) are facial trauma involving the paranasal sinuses. Pneumocephalus is usually asymptomatic and self-limiting. In some cases, however, air accumulation can progress and cause tension pneumocephalus, possibly resulting in permanent neurological damage or even death. Pneumocephalus is a rare complication of positive airway pressure ventilation, and the most common reported cause in such cases is continuous positive airway pressure (CPAP). Pneumocephalus as a complication of positive pressure ventilation is rare in children1–3 and extremely rare in neonates.4 ,5

We report a case of tension pneumocephalus with severe neurological sequelae in a preterm infant receiving respiratory support with high-flow nasal cannula (HFNC) oxygen therapy. The case highlights the importance of the careful use of HFNC therapy, and while pneumocephalus is a rare complication, it needs to be considered, as early diagnosis and appropriate treatment can modify prognosis.

Case report

We report the case of a female infant born to a gravida 4, para 2 mother following an emergency caesarean section due to fetal distress in the 27th week of gestation. Prenatal ultrasound (US) and other examinations had been normal. The infant weighed 710 g at birth and had an Apgar score of 7 and 9 at 1 and 5 min, respectively. She received intermittent CPAP during the first minutes of life, prior to being admitted to the neonatal unit, where she was administered one dose of surfactant and connected to a nasal CPAP (nCPAP) infant flow system through the small-prong nasal interface. At 6 hours of life, a second dose of surfactant was given and the patient was placed on invasive mechanical ventilation until 17 hours of life. Respiratory support was subsequently administered by nasal synchronised intermittent mandatory ventilation (nSIMV) for 24 hours via a Giulia respirator and a nasal cannula with a Smart Flow NIV Kit (extra small RED system). She received HFNC therapy from the second to the fifth days of life to wean her off nSIMV and as alternative to nCPAP. A premature nasal cannula (Fisher & Paykel) and a maximal flow rate of 4 L/min were used. The cannula size was no larger than half the diameter of the infant's nares, and an unretained pacifier was used to keep her mouth closed. Following this, and up to 12 days of life, she received oxygen in the incubator. On day 12, HFNC therapy was restarted as primary respiratory support due to an increase in oxygen requirements. The transfontanellar US performed at birth and repeated at 4 days of life showed no alterations. On the third day of life, while the infant was receiving HFNC respiratory support, a small erosion was observed in the right nostril and columella. The patient was administered empirical antibiotic treatment with ampicillin and gentamicin from birth to 8 days of life; this treatment was then switched to teicoplanin to prevent a central venous access infection. The central venous line and teicoplanin were withdrawn at 11 days of life. Serial analytical tests were normal and acute phase reactants and microbiological cultures were negative. On day 12 of life, the patient presented mild hyperglycaemia (probably related to stress), increased oxygen requirements and respiratory acidosis; none of the clinical or laboratory data were suggestive of bleeding or infection. The transfontanellar US was repeated and an anechoic left parietal area without a midline shift was observed. A 2.5 cm increase in head circumference was observed over the following 4 days (27.5 cm, from −0.05 z-score to 1.06 based on Fenton’s growth chart), with bulging anterior fontanelle and diastasis of the cranial sutures. She remained without antibiotic treatment from day 12 to day 15 and there were no clinical or laboratory changes. Before transfer to our hospital, on day 15, another US scan showed an increase in the anechoic area with a major midline shift. In view of these events, the patient was transferred to our hospital (a tertiary-care centre) from the secondary-care hospital where she was being attended.

Physical examination showed abnormally low sensitivity to stimuli and a large bulging anterior fontanelle with increased tension and considerable diastasis of cranial sutures, in addition to tissue loss and ulceration in the nasal columella. Brain US showed interposition of gas through the anterior fontanelle and a major midline shift to the right, a dilated right lateral ventricle and diffuse white matter hyperechogenicity through the posterior and mastoid fontanelle. Skull X-rays showed pneumocephalus. While there was no skull fracture, a radiolucent line, located in the ethmoid or sphenoid sinuses, was seen running from the nasopharynx to the skull base. A CT scan revealed tension pneumocephalus in the left cerebral hemisphere, midline shifting of up to 5.5 mm, a connection between this midline and the frontal horn of the right lateral ventricle, a hypodensity signal in occipital white matter and the absence of brain parenchyma tissue in the left middle cerebral artery territory; the same hypodense line was observed at the skull base, but we cannot clearly demonstrate the bone defect because the image was low resolution to avoid excessive radiation, given the very young age of the patient; there was no evidence of subcutaneous scalp emphysema or pneumo-orbitis (figure 1). Imaging tests no showed other anomalies suggestive of other causes of pneumocephalus. Evaluation of the nasopharyngeal region showed nasal trauma only. Laboratory tests were repeated and were unremarkable. The diagnosis was tension pneumocephalus induced by HFNC ventilation. The patient was connected to invasive mechanical ventilation and administered empirical antibiotics and analgesia. Intracranial decompression was performed with direct needle puncture through the anterior fontanelle, with an output of 6 mL of air. US was performed periodically and showed a progressive decrease in air content, an increase in intracranial liquid content and a cystic lesion in left basal ganglia, secondary to necrotic changes. Electroencephalogram tracing showed hypoactive and asynchronous bursts of activity. MRI performed at 23 days of life showed marked dilatation of the lateral ventricles with gas in the anterior horns, no midline shifting and absence of the left parietal and frontal brain parenchyma and head of caudate (figure 2). The parents were informed of the severe brain damage and poor prognosis and it was decided to limit the therapeutic effort. The patient died after 3 hours, at the age of 23 days. Autopsy was not performed at the request of the family.

Figure 1

Cranial computed tomography scan: tension pneumocephalus (N) in the left cerebral hemisphere, midline shift connected to the frontal horn of the right lateral ventricle (VD); no visualisation of parenchymal tissue in the territory of the middle cerebral artery and a hypodense line at the base of the skull (arrow).

Figure 2

Brain MRI: marked dilatation of the lateral ventricles with gas in anterior horns (G), without displacement of the midline, absence of visualisation of left parietal and frontal brain parenchyma and the head of the caudate nucleus.


Pneumocephalus is usually caused by head trauma, but other causes of intracranial air include facial tumours that erode through the skull, gas-forming bacterial infections, paranasal sinusitis, otitis, mastoiditis, neurosurgical and facial procedures, and barotrauma.

Pneumocephalus in neonates is very rare. Most of the cases described to date have been associated with complications arising from infections of the central nervous system, such as meningitis due to gas-forming bacteria.

Air leaks (pneumothorax, pneumomediastinum and subcutaneous emphysema) are a well-known complication of positive pressure ventilation, but iatrogenic pneumocephalus in this setting has been reported less frequently. The first case of pneumocephalus associated with nCPAP was reported in 1989 in a 55-year-old woman with sleep apnoea syndrome on nocturnal nCPAP, and a continuous flow of oxygen. Cases of pneumocephalus following positive pressure mask ventilation have also been reported. Pneumocephalus as a complication of the use of nasal cannula or nasopharyngeal catheters is rare. To our knowledge, there have been just three reports of pneumocephalus in paediatric patients receiving respiratory support with oxygen through a nasopharyngeal catheter or a nasal cannula.1–3 Two additional cases identified an association between positive pressure ventilation and pneumoscalp and pneumo-orbitis (but not pneumocephalus) in neonates. In the first case, a premature infant with a gestational age of 30 weeks on mechanical ventilation developed a pneumoscalp after pneumomediastinum and in the second case, a 26-week-old premature infant developed subcutaneous scalp emphysema and pneumo-orbitis while on an HFNC system.4 Only one case of pneumocephalus (associated with subcutaneous scalp emphysema) has been reported in association with the use of nasal cannula in neonates.5 The infant was receiving respiratory support through a low-flow nasal cannula, which was found to have been inserted deeper than planned. Our report is the first of tension pneumocephalus induced by HFNC therapy in the neonatal period.

There have also been reports of other complications during HFNC therapy. Three cases of serious air leak syndrome, with pneumothorax and pneumomediastinum, were reported in older infants and a teenager treated with HFNC in 2013.6

The positive airway pressures generated during HFNC are not usually measured and are unpredictable. The generation of CPAP is influenced by flow rate, nasal cannula size and mouth position. These factors are particularly relevant in preterm infants, since the pressure produced also depends on the infant's weight and it is in the smallest infants receiving the highest flow rates, with their mouth in a fully closed position and a close-fitting nasal cannula that clinically significant levels of airway pressure can be achieved. In our review of the literature, we found large variations for all these factors from one setting to the next. The evidence surrounding pressure generation with HFNC therapy is somewhat conflicting and there have been few reports on the safety and effectiveness of this modality in extremely preterm infants. Nonetheless, despite the lack of clear criteria and guidance, HFNC has become a popular respiratory support option for newborns.

Based on the data available and on our patient's history, we believe that our patient developed pneumocephalus through an open communication with the skull base in relation to positive pressure ventilation in an inherently more fragile area. We believe that the brain damage was secondary to the compressive effect exerted by the air pressure on the brain parenchyma. We cannot establish the exact moment at which damage began, but there was a clear increase in the perimeter of the skull and diastasis of the sutures, in addition to US changes, on day 12, coinciding with the reinstatement of HFNC ventilation, which would have resulted in increased air intake and the development of tension pneumocephalus.

Undiagnosed pneumocephalus can progress to tension pneumocephalus, which represents a true medical emergency and requires direct drainage to prevent permanent neurological damage.

In conclusion, although pneumocephalus is a rare complication of positive pressure ventilation, clinicians need to be aware of this likelihood. A close watch should be kept on any neurological changes (eg, altered level of consciousness or increase in head circumference) in order to avoid diagnostic and therapeutic delays and a fatal outcome. A high index of clinical suspicion is also necessary if nasal trauma (an indirect indicator of excessive pressure) is observed and if laboratory test and transfontanellar US have ruled out other more common causes of acute neurological symptoms, such as intracranial haemorrhage or infections. Because it is non-invasive, US can be repeated as often as necessary and in the case of equivocal findings, other imaging tests can help to guide diagnosis.

We underline the importance of the careful use of positive pressure ventilation systems, especially in premature neonates. It is also important to periodically review insertion depth and to ensure that an adequate airway opening is maintained at the nares. This is achieved by choosing correctly sized prongs that is, prongs that do not completely occlude the nares. Further research is needed to evaluate the safety and effectiveness of HFNC therapy in preterm infants and the frequency of serious adverse events. The resulting knowledge should be used to develop clear guidelines and protocols for the use of HFNC therapy in extremely premature infants.



  • Contributors AI-D took part in patient management, data collection, a critical review of the literature and drafting of the manuscript; AP-M took part in patient management, a critical review of the literature and revision of the manuscript; OL-S and PC took part in patient management; PC, the guarantor of the article, took part in patient management, a critical review of the literature review and manuscript revision. All the authors approved the final version of the manuscript.

  • Competing interests None declared.

  • Patient consent Parental/guardian consent obtained.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • Fantoms
    Ben Stenson