Oesophageal perforation in extreme prematurity

Abstract

Management of oesophageal perforation in extremely premature babies is a challenge and carries a high morbidity. The authors report their experience of three separate cases of oesophageal perforation they encountered over the last 18 months in extremely premature neonates. In the first case, the diagnosis of oesophageal perforation was known in an otherwise stable baby who was treated conservatively with a good outcome. In the second and third cases, the patients proceeded to thoracotomy before the diagnosis of a perforated oesophagus was unveiled. A subsequent diagnosis of trisomy 18 and complex cardiac anomalies in the second case proved incompatible with life and treatment was withdrawn. In the final case, an oesophagostomy was fashioned with a gastrostomy for feeding postoperatively and gastric interposition is planned later this year.

Background

In extremely premature neonates instrumentation to access the upper airway and gastrointestinal tract is an essential part of management. Oesophageal perforation is a rare consequence of this instrumentation but remains most prevalent in this delicate patient group.¹ The following cases illustrate three very different scenarios of oesophageal perforation. The differences that exist between each case demonstrate the variation in presentation, management and unforeseen events which can occur in this entity, supporting the need for increasing awareness and sharing of experiences within this challenging scenario.

Case presentation

Case 1: an extremely premature male neonate of 23+5 weeks was born by spontaneous vaginal delivery at home weighing 640 g. Immediate resuscitation was required by paramedics and subsequent intubation was successful at second attempt on arrival at a neonatal intensive care unit (NICU). The baby required inotropic support for 12 h only. Neonatal co-morbidities included a patent ductus arteriosus, right-sided intraventricular haemorrhage and thrombocytopenia. A naso-gastric tube (NGT) was placed and suction catheters used regularly as part of routine care. On day 7 a blocked endotracheal tube necessitated re-intubation and a subsequent chest radiograph (CXR) demonstrated a pneumomediastinum and NGT entering the right side of his chest. Contrast study (figure 1) confirmed a leak and the patient was transferred to our institution. After resuscitation, we managed the patient conservatively for 4 weeks with total parenteral nutrition, intravenous antibiotics and nil by mouth. On recovery from a septic episode a contrast study in the fifth week confirmed a fully healed oesophagus. The patient resumed feeding and suffered no long-term complications of the oesophageal perforation.

Figure 1.

Chest radiograph following contrast injection into NGT demonstrating leakage of contrast into pleural cavity.

Case 2: An extremely premature female infant, weighing 961 grams was born by emergency caesarean section (EMCS) following a poor cardiotocography trace and abnormal end diastolic flow, at 28+4 weeks gestation at a regional hospital. Worsening polyhydramnios had been noted at 20 and 27 weeks. She had a difficult intubation at birth and a NGT could not be passed. On CXR, the NGT was seen to be within an upper oesophageal pouch and some gas in the stomach. The upper pouch tube was exchanged for a replogle tube and the baby transferred to our unit uneventfully with a diagnosis of oesophageal atresia with a trachea-oesophageal fistula. Initial cardiac assessment was not contributory. At operation, the findings were consistent with oesophageal atresia and distal tracheoesophageal fistula, however the replogle tube was noted in the chest confirming an upper pouch oesophageal perforation. The fistula was ligated and transfixed. Primary anastomosis was initially attempted but abandoned due to a long posterior tear of the upper pouch. The perforation was repaired, both pouches tied off and a gastrostomy formed. Subsequent echocardiography revealed complex congenital heart disease with a double outlet dominant right ventricle and an unbalanced atrioventricular septal defect. Late arrival of the karyotyping results confirmed trisomy 18 as the unifying diagnosis to explain the constellation of anomalies. This combination proved incompatible with life and intensive care treatment was withdrawn on day 12.

Case 3: an extremely premature female neonate, one of sextuplets was born in good condition by EMCS at our institution due to early onset of labour weighing 670 g. Intubation was successful at 4 min on second attempt with a size 2.5 endo-tracheal tube. After 10 min the baby was transferred to NICU in a stable condition. Over the next 15 h two attempts failed to pass an NGT and bloody secretions were noted. The second NGT was secured at point of resistance and flushed with contrast confirming the tip position at the level of the fifth thoracic vertebral body and no contrast seen elsewhere. With bowel gas noted on an abdominal radiograph the working diagnosis of oesophageal atresia with a distal tracheosophageal fistula was confirmed. A right thoracotomy revealed a surprising diagnosis of oesophageal perforation extending from the region of the glottis to the level of the carina in an otherwise normal oesophagus. Primary repair was not possible therefore the distal portion of the oesophagus was tied off, the proximal segment secured as a right cervical oesophagostomy and a gastrostomy fashioned. Surgically the patient remained stable however, at 4 months of age the oesophagostomy became stenosed and required revision. She was discharged home at 6 months of age with a freely draining oesophagostomy and full enteral feeds via the gastrosotomy. She is now 14 months old thriving normally, showing good developmental progression and gastric interposition planned later this year.

Discussion

Iatrogenic oesophageal perforation in the newborn was first described as early as 1961² and has since remained a rare occurrence although the actual incidence is not known. Extreme premature infants are particularly at risk with a death rate of 20–30% following oesophageal perforation.³ Friable mucous membranes in this group are more susceptible to trauma caused by difficult endotracheal and nasogastric intubation or suction catheters.

The most common site of perforation is at the narrowest point at the pharyngoesophageal junction. Three types of injuries have been described by Mollitt et al. First, a pharyngeal pseudo diverticulum is created due to a localised cervical leak; second, a submucosal perforation produces a false passage posterior and parallel to the oesophagus and finally, free spillage of oesophageal contents is seen directly into the pleural cavity.⁴ The different modes of injuries described above explain the wide spectrum of clinical presentations that can result and is reinforced by our three very different cases.

In our first case the diagnosis was suggested following a CXR showing an abnormally positioned NGT. This was later confirmed by a right-sided pneumothorax and a contrast study showing radiopaque material in the pleural cavity (figure 1). This case supports the argument for conservative treatment in an otherwise stable neonate with intravenous antibiotics and total parenteral nutrition.

In the extremely premature neonate if passage of a NGT is difficult it may be repeated under fluoroscopic guidance or avoided until a later date. Bloody secretions should alert the clinician to a potential oesophageal perforation which may be confirmed by contrast study. This raises the question of the use of flexible endoscopy in this situation to allow for early diagnosis at the patient’s bedside before theatre.⁵ This also enables placement of further NGTs under direct vision to allow ongoing enteral feeding without re-entry into the perforation site. Unnecessary thoracotomy could be avoided in the stable patient that could be treated conservatively.⁵ However, this technique could in itself contribute to oesophageal trauma and is technically difficult in such small infants but maybe an area for future review.

In summary, our experience supports the need for selective treatment of this condition.⁴ This is dependent on a number of factors including time of diagnosis, site of perforation, extent of extravasation and the stability of the neonate. An increased awareness in this extremely premature group and supreme care when carrying out instrumentation of such delicate tissue is the most important factor to help avoid this potentially fatal scenario.

Learning points.

• Supreme care during instrumentation of upper airways in extremely premature neonates.

• Conservative treatment of oesophageal perforation is appropriate in selected cases.

• Future development in neonatal endoscopy may assist in diagnosis of oesophageal perforation.