General/Thoracic Surgery: No ownership of primary postoperative haemorrhage following difficult right adrenalectomy
Case Summary:
A 78-year-old male patient had an adrenalectomy for recurrent adrenal haematomata (on a background of myelodysplastic syndrome). During surgery the inferior vena cava (IVC) was injured at the junction of the right hepatic vein, resulting in massive bleeding. The patient lost more than 3 L of blood. The vein was not suture-repaired but controlled with haemostatic agents (TachoSil and Spongostan). Because the mass was adherent to the IVC, part of the mass was left behind and oozing from the area was ‘controlled’. It was noted that there was a 0.5 cm injury to the right diaphragm which was also not repaired. The patient had massive transfusions and was transferred to the intensive care unit (ICU) ventilated and on inotropes.
Initial chest X-ray was satisfactory but on postoperative day 3, chest X-ray showed fluid in the right chest. A computed tomography (CT) scan did not reveal any extravasation of blood from any vessels, and it was felt that this was a pleural effusion, not a haemothorax. An ultrasound-guided pigtail catheter was inserted, which drained haemoserous fluid. Drainage slowed over the following days and the drain was removed on postoperative day 7. At this time, the patient was extubated and transferred to the ward. He was having difficulty swallowing, so a nasogastric tube was inserted. The next day, he became hypotensive, requiring a medical emergency team (MET) call. Chest X-ray showed complete opacification of a right hemithorax. CT with contrast did not reveal any active bleeding (clinicians had suspected bleeding from the site of pigtail insertion). The radiologist reported that, despite being poorly visualised there was a possibility that bleeding was occurring at the site of the IVC injury and tracking into the chest. A fresh haematoma immediately adjacent to the IVC was also reported, and a slow bleed from the IVC or adrenalectomy bed could not be ruled out. There is no evidence to indicate that this report was read or acted upon by any clinician.
The patient had 5 units of red cells, was on inotropes and ventilated. His haemoglobin was 5.0 g/L, which came up to 12.5 g/L with further transfusions. His INR (international normalised ratio) was normal. The primary surgeon contacted the cardiothoracic surgeon, who asked for a chest drain and to monitor the drain output. The thoracic surgeon initially drained 1.7 L of heavily bloodstained fluid, with the drainage eventually slowing. The patient remained stable, but the chest X-ray suggested that he still had a large volume of retained clotted blood in the right chest. The thoracic surgeon was comfortable with progress and opted to monitor, suggesting surgery if further bleeding occurred.
Over the next 2–3 days the patient remained stable, but chest X-ray showed significant residual haemothorax. He had another CT 3 days later, which showed that the fluid had increased mildly with more loculated collections, but the overall appearance was similar. He was offered bronchoscopy, thoracotomy and washout, and decortication. This was done the next day (approximately 2 weeks after first surgery). At operation, bronchoscopy was normal although there was a large volume of organising haemothorax with more than 1 L of blood in the chest. The lung appeared trapped and a decortication was performed. There were no complications and there were no bleeding points. There was no mention of the surgeon examining the diaphragm for any injuries. The surgeon took multiple pleural biopsies and noted that the raw surfaces were oozing. The patient was given further platelets, blood products and haemostatic agents. He was transferred to ICU where the immediate chest X-ray was satisfactory; however, he continued to drain a moderate volume of blood. Chest X-ray in the evening showed a massive haemothorax on the right side. The patient was offered further surgery but the family declined, since the first surgery had not identified any bleeder. He passed away the next day.
Discussion:
There were several issues with the management of this patient.
There was a failure of operative technique in not achieving definitive repair of the IVC injury. Topical haemostasis was reached using agents (TachoSil and Spongostan) contraindicated for major venous bleeding. While a retrohepatic injury can be very difficult to repair, it is not clear from the notes what attempts were made for definitive repair. It is also unclear if the ICU team were made aware that suture repair of the IVC laceration had not been achieved.
Given that the operation was complicated by massive haemorrhage, the postoperative focus should have been on the possibility of ongoing bleeding or rebleeding. This did not occur. Postoperative orders asked for the surgeon to be informed in the event of haemorrhage, but there was no documented explanation for the shock/falling haemoglobin/transfusions observed during the first 2 postoperative days. By day 2, primary postoperative haemorrhage (likely from the IVC) should have been considered by the primary surgeon. Tachycardia, hypotension, the need for further transfusion, and queries by ICU staff regarding the possibility of postoperative ooze, all indicated that rebleeding had occurred. It is disappointing that there was no mention of the need for a CT or reoperation by the primary surgeon in order to address this. Ideally, haemorrhage should have been excluded as early as the first night. Falling haemoglobin (first ICU haemoglobin of 88 g/L had dropped to 70 g/L the next morning) is a classical feature of haemorrhagic shock. Instead, he was managed with vasopressors and transfusion.
The patient died of postoperative haemorrhage following a difficult right subtotal adrenalectomy complicated by IVC injury. He bled soon after arriving in ICU but this went unrecognised despite the need for vasopressors and a transfusion. He bled again on days 8–9 and decompensated on day 10. The surgical team faded to the periphery. Management was at odds with CT reports repeatedly suggesting that bleeding was tracking up into the right chest from the surgical site. The patient exsanguinated after the haemothorax was decompressed.
Clinical Lessons:
It is important to read all the case notes, including the operation notes and specialists reports, properly and in detail.
The diaphragmatic injury was documented by the first surgeon, who stated that it was ‘closed’ off by the liver. While the liver will prevent solid organ herniation it will not prevent any fluid, in this case blood, from entering the chest. The radiology reports from the day of the initial massive bleed clearly state that there was a new haematoma in the adrenal bed as well haematoma around the IVC tracking into the chest. This clearly pointed to fresh bleeding in the abdomen; however, no clinicians considered this possibility, nor did they consider the bleeding in the abdomen with injury to the diaphragm as the reason for the haemothorax.
Most likely the large haemothorax had tamponaded further bleeds due to the positive intrathoracic pressure. With decortication, the fibrin and organising tissue were removed from the diaphragm. Manipulation of the clots and fibrin on the diaphragm could have dislodged any clots around the infra-diaphragmatic IVC, precipitating further bleeding, which likely tracked up into the chest again. The only other explanation would be blood oozing from the raw areas after decortication, which seems unlikely considering there were no reports of any active bleeding during the operation and haemostatic agents were used freely immediately after surgery.
The patient’s only chance of survival was if the IVC injury had been suture-repaired with the help of vascular surgeons. There is no substitute for proper control of haemorrhage from major vessels. While the patient was hypovolemic the bleeding stopped, only to recur once venous pressure normalised. While it cannot be conclusively stated that the IVC injury was the source of the bleed, the only other possible source was the remaining capsule of the adrenal mass, which was noted to be oozing.
Thoracic surgery was needed for the organising haemothorax, but the surgeon seems to have been unaware of the radiology report or the surgeon’s report of the diaphragmatic injury. The surgeon should have been aware of these reports and consulted vascular colleagues to develop an action plan to repair the diaphragmatic injury and suspected intra-abdominal source of bleeding. An earlier operation would have been preferable, when surgical repair would have been easier, although it is doubtful that this would have affected the outcome for this patient. The rationale for a bronchoscopy in this patient is unclear.
For a patient with coagulation problems, it is wise to use video-assisted thoracoscopic surgery (VATS) to minimise the extent of dissection and decortication, so blood loss can be minimised. VATS and irrigation with minimal decortication would have expanded the lung adequately to fill the space, considering it was less than 2 weeks old. In an elderly patient it may be better to accept a sub-optimal inflation of the lung rather than risk further bleeding.
References:
Chou YP, Lin HL, Wu TC. Video-assisted thoracoscopic surgery for retained hemothorax in blunt chest trauma. Curr Opin Pulm Med. 2015;21(4):393-8.
Godat L, Cantrell E, Coimbra R. Thoracoscopic Management of Traumatic Sequelae. Current Trauma Reports. 2016;2(3):144-50.
Disclaimer:
Please note that these cases are edited from ANZASM first- or second-line assessments that have been generated by expert surgeons in the field. Any recommendations relate to these cases as they were presented.
|
