Iatrogenic tracheobronchial tear during paediatric thoracoscopic oesophageal repair: an anaesthetic challenge

Abstract

Iatrogenic tracheobronchial injury is rare. Limited data is available on such injuries in infants and management of these is challenging. We present a 7-month-old male infant who was diagnosed with oesophageal stricture, and was planned for thoracoscopic repair of the same. Anaesthetic management with a single lumen tracheal tube ensued. However, an intraoperative loss of capnogram and desaturation led to the diagnosis of iatrogenic tracheobronchial injury. After unsuccessful pursuits with a Fogarty catheter through the tracheobronchial tear and through the right bronchus for lung isolation, blind left endobronchial intubation with a smaller tube and one lung ventilation was successful. The rent was repaired by an open procedure, and oesophageal surgery completed. However, the child succumbed and died of multiorgan failure 3 days later.

Background

Iatrogenic tracheal injury in the paediatric age group is a rare entity. However, it can occur most commonly during traumatic tracheal intubations¹ and, to a lesser extent, with thoracoscopic surgeries, tracheostomy or bronchoscopy. The clinical presentation depends on the degree of injury and can range from asymptomatic to cervicothoracic subcutaneous emphysema, cyanosis, respiratory distress and pneumothorax secondary to air leakage from the tear.^{2 3} The consequences of a tracheal tear can be fatal, even more so in the paediatric population. However, there is limited data available on this in infants, thus making its management challenging.

Case presentation

We present a 7-month-old male infant, weighing 7 kg, a known case of type-C tracheo-oesophageal fistula diagnosed at birth, status post open thoracotomy and fistula repair and multiple oesophageal dilatations, for thoracoscopic repair of oesophageal stricture. He was also a known case of congenital heart disease with a 4 mm ostium secondum atrial septal defect with left-to-right shunt and small patent ductus arteriosus. His general physical and systemical examination, and blood work-up were within normal limits.

On the day of the surgery, the infant was shifted to the operation theatre, a working intravenous line was secured and the infant was started on 1% dextrose-ringer lactate using a burette. The indication for one-lung ventilation (OLV) in this child was purely a surgical one, and hence, following a discussion with the paediatric surgeons and after weighing the benefit-risk ratio, it was decided that lung isolation was not essential for the procedure, due to the experience and comfort of the surgeons with video-assisted thoracic surgeries (VATS) successfully with two-lung ventilation (TLV), and hence we proceeded with TLV. Standard American Society of Anaesthesiologists (ASA) monitors were connected and baseline vitals recorded. A convective body warmer and fluid warmer, along with warmed cotton sheets and cap was used to maintain normothermia. Anaesthesia was induced with 14 mg intravenous propofol and 10 mcg intravenous fentanyl. After confirming bag-mask ventilation, neuromuscular blockade was achieved with 3.5 mg intravenous atracurium and anaesthesia was deepened with 4% sevoflurane in oxygen to achieve a minimum anaesthetic concentration (MAC) of 1.7. The infant was intubated with a 3.0 mm internal diameter (ID) micro-cuffed (Kimberly-Clark) endotracheal tube (ETT), obtaining grade 1 Cormack-Lehane laryngoscopic view. Bilateral air entry was confirmed and the tube was fixed at 10 cm depth to the right angle of the mouth. Pressure control ventilation was initiated. A 10 Fr nasogastric tube was inserted into the stomach and fixed. The child was positioned for surgery in the left semi-prone position, pressure points padded and tube position reconfirmed by auscultation. Partial right lung collapse was achieved with carbon dioxide (CO₂) pneumothorax at 5 mm Hg.

Approximately 45 min into the surgery, during mobilisation of the oesophagus from the trachea, there was a sudden drop in end-tidal carbon dioxide (ETCO₂). We were unable to ventilate the patient with adequate tidal volumes. Accidental displacement of the ETT out of the trachea and other common causes were ruled out, and the incident was immediately brought to the attention of the surgeons. They suspected a tear in the trachea secondary to surgical dissection, attributable to adhesions in the surgical field due to chronic inflammation and previous surgical scar tissue. At this point, there was a decrease in blood pressure as well. The ETT was pushed in further gently to bypass the suspected tracheal rent and we were able to ventilate the infant with this manoeuvre. Auscultation revealed bilateral air entry but with conducted sounds on both sides. The capnogram would be seen only when pressure was applied in the surgical field through the scope, and when this pressure was released the ETCO₂ would reduce to 0 mm Hg, indicating that ventilation was possible only transiently when the tracheal tear was occluded with the thoracoscopic instruments. Keeping in mind the possibility of lung parenchymal injury leading to obstruction of the ETT with blood, sterile suctioning of the tube was done and blood-tinged secretions were aspirated. Saturation remained at 85% to 90% using these methods for the next 45 min, with ventilation being possible intermittently. The surgeons were unable to track the exact extent and location of the tear in the trachea during this period. Thereafter, there was flattening of the ETCO₂ trace, and ventilation was hampered. Patient began desaturating, was turned supine and cardiopulmonary resuscitation started as per advanced cardiac life support guidelines, in view of heart rate <60 beats/min with signs of poor perfusion, achieving return of spontaneous circulation (ROSC) after three cycles. The ETT was removed due to suspicion of repeated tube block. The trachea was reintubated with a 3.0 mm ID micro-cuffed ETT. In spite of ROSC, saturation remained on the lower side, even with 100% oxygen. Hence, we performed a fibreoptic bronchoscopy (FOB) to see if the tracheal tear can be visualised from within the trachea. To facilitate passage of the bronchoscope through the ETT, infant was reintubated with a 4.0 mm ID ETT. Through the FOB, the tracheal rent could not be positively identified, and therefore, the cause of arrest and intermittent loss of ventilation remained unclear. A possibility of right pneumothorax was suspected, secondary to the rent in the tracheobronchial tree and a right intercostal drain was placed, but with no improvement.

The child was repositioned in the semi-prone position for thoracoscopic identification of the suspected rent, as seen in figure 1A. An attempt to pass 5 Fr Fogarty catheter through the suspected rent in the right bronchial tree (thoracoscopically) to isolate the right lung was tried in vain (figure 1B). Using a FOB through the ETT, the tip of the Fogarty catheter was visualised confirming that the rent in the bronchial tree had been correctly detected, and in spite of initial correct placement of the Fogarty catheter, there was no improvement in ventilation due to slipping out of the Fogarty catheter. Ventilating the infant and maintaining peripheral capillary oxygen saturation (SpO₂) >90% was extremely difficult during this period. An effort to pass another 5 Fr Fogarty catheter through the ETT was done to isolate the right lung but was unsuccessful. The child was given 100% oxygen throughout and was maintaining saturation of 85% to 87%. Preparation for high-frequency oscillatory ventilation and a cardiothoracic opinion regarding cardiopulmonary bypass was sought. After discussion with the surgical team, a decision to abandon thoracoscopy and proceed with a thoracotomy and repair of the defect was made. To aid the ease of surgical repair, left endobronchial intubation with 3.0 mm ID ETT, with the help of FOB was attempted but was unsuccessful due to incompatible size of the ETT. A final attempt at left one-lung ventilation with a 2.5 mm un-cuffed ETT, using blind technique and rotating the head to the right side was successful. With this, we were able to ventilate the left lung with adequate tidal volume and the saturation picked up to 99% with ETCO₂ between 45 to 50 mm Hg. After the rent was surgically repaired, the child was re-intubated with a 3.5 mm un-cuffed ETT for two-lung ventilation and repair of the oesophageal stricture. The remainder of the surgery was uneventful and the child was shifted to the paediatric intensive care unit with ETT in-situ for further postoperative management. Figure 2 shows a comparison of the preoperative and postoperative chest X-ray, showing obvious collapse of the right lung with pneumothorax. A CT chest done in the postoperative period confirmed right pneumothorax, and revealed consolidation and collapse of the underlying lung, with right bronchial narrowing (figure 3). The infant expired after 3 days succumbing to sepsis and multiorgan failure.

Figure 1.

A. Detection of tracheobronchial tear; B. Insertion of Fogarty catheter thoracoscopically through the rent.

Figure 2.

A. Preoperative CXR; B. Postoperative CXR. CXR, chest X-ray.

Figure 3.

Postoperative CT thorax showing consolidatsion, ground glass opacities, pneumothorax and right pleural effusion with narrow right main bronchus.

Outcome and follow-up

The infant expired after 3 days succumbing to sepsis and multiorgan failure.

Discussion

Tracheal injury is a very rare complication of thoracoscopic surgeries, and there is insufficient evidence regarding standard management of these cases. Kim et al have suggested that balloon dilation for benign tracheal strictures can be attempted as although the incidence of secondary tracheobronchial lacerations is 51.6%, they are mostly superficial and very rarely progress to transmural injuries.⁴ Dwivedi et al report a case of iatrogenic rupture of carina during repair of type-C tracheo-oesophageal fistula in a preterm infant, which was managed by intra-thoracic endobronchial insertion of the ETT by the surgeon, until further repair of the rent.⁵ In another case, Kwok et al reported a dramatical increase in the ETCO₂ after a small accidental tear of the main stem bronchus during thoracoscopic repair of tracheo-oesophageal fistula, due to uptake of the gas from the capnothorax. After conversion to thoracotomy, surgical repair of the rent was done with intermittent periods of apnoea.⁶ Dango et al report a case of post-traumatic laceration of main stem bronchus with digital repositioning of the ETT during surgery for optimal ventilation.⁷ Lahori et al encountered an iatrogenic tracheal tear during rigid-bronchoscopic retrieval of a foreign body in a 7-year-old. Due to unavailability of FOB, the rigid bronchoscope was used as a conduit for an airway exchange catheter for endobronchial intubation.⁸ Ho et al report a case of emergency blind left main stem bronchial intubation in a neonate with rupture of emphysematous bullae on the right side, with a similar technique of turning the head to the right, and rotation of the ETT by 180⁰, such that the Murphy’s eye of the ETT faces left instead of right. The knowledge of this technique adds to the paediatric anaesthesiologist’s armamentarium, and can be invaluable in the absence of FOB, bronchial blockers and in emergency situations.⁹ Al Adawy et al have compared respiratory and haemodynamical parameters with OLV and TLV during VATS, and have not found any significant differences between the groups. As the chest wall is more compliant in infants, there is no significant increase in airway pressures or lung compliance. Although VATS is not an absolute indication for OLV, they have found that the visualisation of the surgical field was better with OLV. Hence, they have concluded that OLV can be considered as an alternative to TLV for VATS.¹⁰

In the present case, one would wonder whether lung isolation from the start could have been done with bronchial blocker, and whether that would have averted the injury. On the contrary, in this infant, lung isolation would have made the injury go unnoticed intraoperatively. The isolated uninjured lung would get ventilated unhindered until the end of the surgery without revealing the tear in the tracheobronchial tree. The injury would probably get detected only at the end of the surgery when two-lung ventilation would be resumed. However, in our case, the tracheobronchial tear was picked up intraoperatively due to loss of capnogram. The diagnosis of bronchial tear was one of exclusion and was not very clearly seen, both thoracoscopically and via FOB, in the beginning. Only after the common causes for loss of capnogram and ventilation were ruled out, iatrogenic injury was suspected. Even detection of the exact location of the tear was difficult and time-consuming. The sequence of measures taken to temporarily seal the rent to facilitate ventilation was done under the pressure of loss of ventilation, desaturation and even a transient cardiac arrest. Probably, early conversion of thoracoscopic procedure into an open procedure could have limited the physiological insults to the infant.

Learning points.

• Iatrogenic tracheobronchial tear is a rare and challenging complication of paediatric thoracoscopic surgeries, and diagnosing this can be difficult.

• High degree of suspicion and early detection is of key importance.

• Intraoperative lung isolation in paediatrics, especially in an emergency situation, may be extremely difficult due to narrow airway calibre and airway oedema.

• A low threshold has to be kept for converting to open thoracotomy when complications are encountered.