Autologous Blood Patch and Lung Ultrasound for Persistent Air Leak in an Extremely Premature Neonate

Pneumothorax is a well-known complication of lung disease in the neonate, with an incidence of 10% in extremely preterm infants, though much higher in the setting of preterm, premature, rupture of membranes (PPROM) with oligohydramnios. Additionally, Infants who experience an air leak requiring an intercostal chest drain have a higher rate of mortality than those who do not experience an air leak [1]. It is rare to have a persistent air leak beyond 7 days, and there is a paucity of literature on treatment and prognosis, though it is generally accepted to be a complication associated with high mortality. This case presentation outlines a novel approach to achieving resolution of a persistent air leak in an extremely premature neonate with the use of an autologous blood patch.

The patient is a female infant with pulmonary hypoplasia and respiratory distress syndrome who experienced persistent air leak for 28 days. She was born at 24 + 4 weeks gestation with a birth weight of 697 g. Her mother's pregnancy was high risk and complicated by placenta increta along with PPROM from 18 weeks gestation, with associated anhydramnios. She had a planned high-risk caesarean section delivery at 24 weeks after a full course of steroids and magnesium sulphate. This was a complicated delivery with extensive intraoperative blood loss estimated at 16 L, controlled with surgical and radiological intervention along with activation of a massive transfusion protocol.

The infant required intubation at birth. She had evidence of severe Respiratory Distress Syndrome and Pulmonary Hypoplasia. She was transferred to the Neonatal Intensive Care and placed on high-frequency oscillatory ventilation (HFOV); settings to achieve preductal saturations of 90%–95% and PaCO₂ 45–60 mmHg were mean airway pressure (MAP) 15, Hertz of 15, amplitude of 50, fraction of inspired oxygen (FiO₂) 1.0. She received a total of four doses of surfactant (Poractant alpha) over the first 48 h due to the ongoing severity of her lung disease, air leaks, and high-pressure ventilation requirement. At 4 h of age, she developed a right-sided tension pneumothorax requiring intercostal chest drain insertion (Figure 1). At 30 h, she developed a left-sided tension pneumothorax requiring intercostal chest drain insertion (Figure 2). Over the following 28 days, she experienced persistent right-sided air leak. Her treatment included multiple chest drains, moderate-dose systemic steroids (3.6 mg/kg cumulative dose of Dexamethasone) and attempts to manage on non-invasive ventilation (triggered non invasive positive pressure ventilation and nasal high frequency ventilation). Chest drains were directed anteriorly (confirmed with a lateral chest x-ray) and attached to 20 cm H₂O negative pressure suction. They were observed for patency (swinging) and were always effective at decompressing the pneumothorax. With resolution of bubbling, each attempt at coming off suction or removing the chest drains led to reaccumulation on the right side. At 28 days of age, she remained dependent on conventional mechanical ventilation and right-sided intercostal chest drain. In her favour was her excellent nutrition, achieving full enteral feeds of mother's expressed breast milk (fortified) by day 16 of life and her normal screening head ultrasounds.

At 28 days of age, an alternative treatment plan was developed with the family. She was transitioned back to high-frequency oscillatory ventilation with moderate sedation (Morphine and Dexmedetomidine), a right-sided autologous blood patch was performed, and she was nursed with extended periods of time prone, right side down, or skin to skin on her parents, with minimal handling. The sterile autologous blood patch procedure involved sampling 1.6 mL (2 mL/kg) whole venous blood from the patient and immediately injecting it down her right intercostal chest drain into the pleural space. The catheter was cleared with a flush of 2 mL of 0.9% Sodium Chloride, followed by 3 mL of air. The underwater seal drain was left unclamped and raised 50 cm above the patient, with the suction turned off. No further bubbling was observed from the underwater seal drain. Regular point-of-care lung ultrasound surveillance over the following hours confirmed bilateral lung sliding with no re-accumulation of air (Figure 3). Significant right-sided basal and apical atelectasis was observed on ultrasound, but no attempt was made to recruit these areas; instead, a high fraction of oxygen was tolerated. The intercostal chest drain was successfully removed 72 h later with no re-accumulation of pneumothorax. Haemoglobin and haematocrit prior to the procedure were 109 g/L and 33.5%, respectively; following the procedure, they were 97 g/L and 30%.

Baby A was discharged home at three and a half months corrected age, on low flow oxygen during the day and humidified high flow nasal cannula at night time.

Premature infants with hypoplastic lungs secondary to PPROM have a high mortality [2]. Persistent air leak, defined as ongoing pneumothorax beyond 7 days, is a challenging complication to manage in these patients and contributes to the high mortality rate [2, 3]. The commonly used approach involves treating the underlying lung condition, the use of multiple intercostal chest drains to decompress the air leak and achieve pleural apposition, and lung protective ventilation strategies [4]. Larger infants may progress to thoracotomy and direct visualisation and sealing of the persistent pleural air leak. For infants that are too small for thoracotomy, there is scant literature describing alternative options. Case reports include approaches such as selective bronchial intubation, pleurodesis with sclerosing agents, and newer approaches with the use of fibrin glue [5-7]. This case describes a novel approach to the treatment of neonatal persistent air leak with an autologous blood patch.

Autologous blood patch has been described in adults and children as an effective therapeutic option [3]. A small case series by Andrade-Montesdeoca describes the successful outcome in three neonates [8], no published cases are available of treatment with autologous blood patch in extremely premature neonates. In our case, the parents described the concept of using their infant's own blood as more acceptable compared to the instillation of a foreign or toxic substance. It was also cost-effective compared to the off-label use of fibrin glue.

The other management strategy worthy of discussion is the use of point of care lung ultrasound [9]. Lung Ultrasound was invaluable in this case given the risk of life-threatening re-accumulation of a tension pneumothorax following the procedure. This was a significant risk as there was a high chance of drain occlusion with a blood clot. With lung ultrasound, real-time pleural apposition was able to be continuously monitored following the autologous blood patch procedure, thus mitigating the risk of unrecognised, life-threatening tension pneumothorax. It is important to mention a major limitation of lung USS in the assessment of pneumothorax is that the size and depth of the air collection can not be quantified; hence, the importance of clinical assessment and consideration of a chest x-ray.

Lung ultrasound gave us the confidence to persevere with the procedure despite the very high fraction of oxygen, as we were able to see that the high oxygen requirement was due to extensive atelectasis rather than pneumothorax. Without this real-time information, the procedure would likely have been abandoned early. The other benefit of lung ultrasound imaging compared to traditional radiography was the ability to image the infant's chest while she was prone and skin to skin with her parent. Traditional radiography involves significant handling of the infant, which in turn has detrimental effects on ventilation and potential worsening of the VQ mismatch and air leak. An interesting observation from the follow-up USS was that there was still lung sliding occurring 72 h after the successful autologous blood patch. Lung sliding would not occur in the setting of pleural adhesion, which may suggest that there is an alternative mechanism of effect from autologous blood patch that leads to sealing of the air leak.

There are some limitations to this technique. The blood sampling required is significant, though this case demonstrated minimal effect on the haemoglobin and haematocrit. Lung ultrasound is an evolving skill in Neonatal intensive care, and may not always be available to support such a procedure. The risk of an obstructed chest tube may be a valid concern, which limits the uptake of this procedure.

I have written informed consent from the parents of baby AO for submission of this case report. Ethics approval was not sought for a single retrospective case report as per institutional practice.

The author declares no conflicts of interest.