Delayed Diagnosis and Treatment of an Occult Hemothorax Following Complicated Central Line Insertion Leads to Cardiac Arrest
A 2-year-old girl with acute myelogenous leukemia and thrombocytopenia (platelet count 26,000 per microliter) with marked generalized petechiae and bruising was scheduled for implantation of a central venous catheter with a subcutaneous port. After reviewing the patient's electronic health records but without seeing the patient, who was in another hospital, the anesthetist asked the surgeon to order a platelet transfusion to increase the child’s platelet count to above 50,000 per microliter. The patient was transported to the operating room while still receiving the platelet transfusion, so a follow-up platelet count had not yet been obtained.
After induction of general anesthesia, the assistant surgeon tried 10 times to cannulate either subclavian vein without using ultrasound guidance. He finally succeeded in cannulating the left subclavian vein and, after fluoroscopically confirming that the tip of the catheter was in the lower half of the superior vena cava, he closed the wound. The patient was then awakened from general anesthesia and transferred to the post-anesthesia care unit. The surgeons had not noticed a moderate left-sided hemothorax on the post-procedure chest radiograph. This finding was reported by the on-duty radiologist but not conveyed to the surgeons. The anesthesiologist, who remained in the operating room, was called to see the patient who was pale, crying inconsolably, with fluctuating blood pressure. Realizing the seriousness of the child’s bleeding condition, he alerted the patient's surgeons and the pediatric intensive care physician, but they were unable to arrive at the child’s bedside for 30 minutes and 45 minutes, respectively, so the anesthesiologist ordered blood for an urgent transfusion. Before the child could be transfused, she suddenly collapsed and developed cardiac arrest accompanied by pulseless electrical activity (PEA). A “code blue” was called and the child was successfully resuscitated after about 20 minutes, following insertion of a thoracostomy drainage (chest) tube.
Unfortunately, the surgeon had damaged an intercostal artery when he inserted the chest tube emergently, which caused further bleeding and two additional episodes of PEA arrest. The patient required bedside thoracotomy and a prolonged pediatric intensive care unit (PICU) stay, during which she developed a subdural hematoma and ischemic encephalopathy.
By Gary Raff, MD, and Brian Goudy, MD
This case highlights a cascade of preventable iatrogenic events after a common procedure that is generally considered low-risk. The delay in diagnosis and subsequent treatment of a hemothorax resulted in a preventable and devastating series of events. Both errors in judgment and errors in technique occurred, and the attending surgeon was ultimately responsible for the outcome. One of the primary roles of the surgical team is to mitigate risk once the risks and benefits of the planned procedure have been assessed and the surgeon agrees to perform the procedure. For central line insertions, the risks include initial catheter malposition and bleeding. The proceduralist must also be on the lookout for injury to surrounding structures (such as arteries, other veins, nerves, thoracic duct, lung, trachea, etc.) and perforations of the vein being accessed or large central veins or heart. If the catheter is malpositioned, it can contribute to delayed complications such as erosion of the catheter through the vein wall or heart, vessel thrombosis, vessel stenosis, and infection (such as central line-associated blood stream infection [CLABSI] or endocarditis). Occlusion of the central line and embolization of thrombus or air from careless flushing of the catheter are also known risks, especially in children with congenital heart disease and complete mixing lesions.1,2
In this case, the patient was thrombocytopenic with clinical evidence of coagulopathy. Appropriately, a platelet transfusion was ordered and was running while the patient was on the way to the operating room. A follow-up platelet count would have provided some assurance that the procedure could be performed safely, with a manageable risk of serious bleeding.
The next issue the surgeon had to address was choosing a surgical approach that would be most appropriate for this patient. In patients at serious risk for bleeding after central line placement, peripheral cannulation is a reasonable alternative to subclavian vein cannulation. Indeed, Kanji et al. showed that in patients undergoing extracorporeal membrane oxygenation, peripheral cannulation was associated with a marked reduction in blood product utilization (i.e., packed red blood cells, platelets, fresh frozen plasma, and cryoprecipitate transfusions) as well as decreased incidence of bleeding from the cannulation site (18% vs 64%).3 Although the patient population in that study had a different indication for central venous access than the patient in the current case, one can reasonably infer that peripheral cannulation would also be safer in patients with similar preexisting coagulopathy (such as a thrombocytopenic patient with a hematologic malignancy). In patients such as the two-year-old in this case, the risk of bleeding into the pleural space must be minimized as such bleeding is difficult to detect until there is a hemodynamic embarrassment, compared with bleeding outside the pleural space, such as in a subcutaneous tunnel. A facial vein cutdown or external jugular vein cutdown would have been safer approaches in this case, as these procedures mitigate the risk of bleeding and pneumothorax compared with the typical subclavian vein or internal jugular vein approach.1 Implanting chemotherapy ports via the external jugular vein is a reliable and safe approach in patients requiring long-term chemotherapy.4 Furthermore, in a retrospective analysis of pediatric patients, Bawazier et al. showed that Hickman central venous catheters inserted via open cutdown of the external jugular tended to have fewer complications (although not statistically significantly) than those inserted percutaneously.5
Despite the potential advantages of peripheral cannulation, a central approach is often required, and the internal jugular vein or subclavian vein can be used. In these situations, however, it is very important that surgeons use all tools at their disposal to mitigate the accompanying risks. In cases involving percutaneous cannulation of the subclavian vein, the use of ultrasound guidance or fluoroscopy would be recommended for high-risk patients, and especially for patients with thrombocytopenia, due to the decreased risk of lung injury as compared with reliance on external anatomic landmarks. In a prospective observational study, Sidoti et al. showed a higher rate of first pass success, fewer attempts, and overall decreased complication rates (the most common being arterial puncture) when ultrasound was used for subclavian vein cannulation versus landmark guided approaches.6 A micropuncture set that includes a 22g needle and 0.018 inch coaxial wire is useful for reducing bleeding if the artery is inadvertently entered. In small children, it is best to avoid J-tipped wires as the radius of the curve is often larger than the size of the vein being cannulated; instead, a very soft-tipped straight wire is preferred for access.1
Even with the ultrasound or fluoroscopic guidance, which was not used in this case, attempting bilateral subclavian vein cannulation is not appropriate in children. In addition, it is not appropriate for resident physicians or other trainees to perform as acting primary surgeon if they are not fully trained and prepared to handle technical problems that may arise during a procedure. Certainly, allowing a trainee so many attempts to cannulate the subclavian vein in the patient in this case was a disservice to the patient and, arguably, to the trainee as well. The trainee apparently lacked the appropriate experience, tools, and guidance from the attending surgeon that were needed to perform this procedure safely. In numerous studies of central venous cannulation in children and adults, the experience of the proceduralist and a low number of cannulation attempts have been found to be important in achieving successful placement and reducing complications.7-9 Although there is no formal guidance from surgical or medical professional societies on when to permit trainees to attempt procedures and how many attempts to allow, in a pediatric patient with coagulopathy, it seems unwise to allow more than 1 or 2 attempts to cannulate by a trainee. Although simulation is a useful method for teaching procedural skills, it does not substitute for closely supervised experience with a substantial number of patients of various ages and sizes.
Perhaps equally troubling as the excessive number of cannulation attempts in this case is the delay in identifying a known serious complication of subclavian vein cannulation. Multiple non-radiographically guided attempts to cannulate the subclavian vein in a thrombocytopenic patient led to a high risk for complications such as hemothorax and pneumothorax. It is concerning that the diagnosis of hemothorax was delayed despite the availability of radiographic evidence demonstrating its presence as well as a radiologist confirming the diagnosis. More timely identification of this complication, with appropriate intervention (which could have been undertaken under better conditions than during active cardiopulmonary resuscitation), would have likely prevented the patient’s cardiac arrest and resultant hypoxic brain injury. It is the responsibility of the proceduralist to look for evidence of early complications associated with catheter placement such as catheter malposition, injury to adjacent structures, or perforation of the vessel that was cannulated. Early identification of complications and timely intervention to mitigate them are central to surgical practice no matter the operation performed. The importance of this topic is such that surgical morbidity and mortality conferences often include sessions on delayed diagnosis or identification of complications.
Additionally, in this case, the surgeon damaged an intercostal artery while inserting the chest tube that was needed to address the patient’s iatrogenic hemothorax. This too was an easily preventable complication, given appropriate training and supervision by an attending physician. In this patient, an injury to the intercostal artery suggests that the technique utilized to insert the chest tube was problematic and that appropriate landmarks were not identified, presumably related to the need to place the chest tube during a cardiac arrest. If the hemothorax had been promptly identified, then the chest tube would have been placed under more controlled conditions. All operations have increased risk when done emergently. This is the rationale behind the “E” addition to the ASA physical status classification of patients by the American Society of Anesthesiologists.9 Injury to the intercostal artery is a rare but life-threatening complication associated with thoracostomy.10,11 The intercostal artery runs inferior to its associated rib, within the costal groove. Standard well-accepted thoracostomy technique involves dissection and tube placement immediately superior to the rib to avoid the intercoastal neurovascular bundle.12 Although there is some intercostal artery variability and exposure in the intercostal space, this anatomic variation is associated with more elderly patients and more cephalad rib spaces.13,14 Furthermore, incision or needle insertion should be made in the safety triangle—the area bounded anteriorly by pectoralis major, posteriorly by the latissimus dorsi, inferiorly by the fifth intercostal space, and superiorly by the axilla—to avoid major structures and to minimize complications. More posterior approaches increase the risk of intercostal artery injury.12 Although the method of thoracostomy used in this case is unknown, the occurrence of such a rare pediatric complication suggests there was likely a deficiency in the technique used.
Approach to Improving Patient Safety
Errors in both judgement and technique led to this adverse outcome. Because there is usually more than one way to approach and perform an operation, it is up to the surgeon to be experienced in as many of these as possible to be able to select and safely perform the approach that is safest for a particular patient or circumstance. In this case, a cutdown should have been used to access the internal jugular, facial, or external jugular vein, or ultrasound guidance should have been used to help minimize the number of subclavian vein cannulation attempts. In addition, the number of cannulation attempts by the trainee should have been limited, and care providers should have exercised more vigilance with regard to the possibility of serious complications. Prompt review of the post-procedure chest x-ray should have been carried out, ideally by the operator but, if unavailable, by a physician qualified to identify a hemothorax or pneumothorax and place a chest drain promptly under controlled conditions or immediately communicate the findings to the proceduralist. Many hospitals, including ours, have a priority system for communicating radiologic finding and systems in place to help prevent delays in diagnosis and treatment, If a life-threatening condition such as pneumothorax or hemothorax is identified on chest x-ray, the care team is immediately contacted by the radiologist. Having such systems in place to rapidly identify complications, and communicate them to the care team in a timely manner, are central to good patient care. A preoperative discussion between the attending surgeon and the trainee regarding anticipated risks and strategies to reduce them, including shared understandings about the allowable number and location of cannulation attempts, could also have improved the outcome in this case. At such a pre-procedure pause, the surgeon has an opportunity to communicate anticipated risks of the procedure, risks that are specific to that procedure and for that patient, to the team. A debrief that includes the entire operating team should also be performed after the procedure to review how the procedure went and to identify any system issues or potential areas of improvement. In this case, review of the numerous failed attempts at cannulation should have raised concern about possible injury to surrounding structures and whether there had been catheter malposition.
Adverse outcomes of the type that occurred in this case are preventable and prevention begins with good judgement, good skills, and good technique. In the hospital’s root cause analysis, it would be important to examine the behaviors and actions of the participants, and especially members of the surgical team, with attention to the possibilities of impairment and poor training.15 In a training institution, best practices include providing clear guidance to trainees about the limits of their skills and experience, ensuring ongoing and active supervision in high-risk settings, and informing and encouraging trainees to seek immediate input from more experienced physicians as soon as questions or problems arise. Following these best practices could have resulted in a much better outcome for the patient in this case.
- Assessing the risk versus benefit of doing the procedure is not enough. The way a procedure is performed as well as mitigation of risk during the procedure must be thought out before the procedure is performed.
- Always have a low threshold to look for complications from any procedure.
- The attending must balance the education of proceduralist trainees with risk to the patient.
- Prompt review of diagnostic studies that identify serious complications must be done by a qualified individual.
- A debrief should be performed after every procedure with an eye towards identifying potential complications and overall improvement.
Gary Raff, MD
Chief, Division of Pediatric Cardiothoracic Surgery
Co-Chair, Children’s Hospital Multidisciplinary Continuous Quality Assurance Committee
UC Davis Health
Brian Goudy, MD
Assistant Professor, Critical Care Medicine
Department of Pediatrics
UC Davis Children's Hospital
- Church JT, Jarboe MD. Vascular access in the pediatric population. Surg Clin North Am. 2017;97(1):113-128. [Available at]
- K Kornbau C, Lee KC, Hughes GD, Firstenberg MS. Central line complications. Int J Crit Illn Inj Sci. 2015;5(3):170-178. [Free full text]
- Kanji HD, Schulze CJ, Oreopoulos A, Lehr EJ, Wang W, MacArthur RM. Peripheral versus central cannulation for extracorporeal membrane oxygenation: a comparison of limb ischemia and transfusion requirements. Thorac Cardiovasc Surg. 2010;58(8):459-462. [Available at]
- Mehta N, Desai SM, Dhakad V, Patel D, Saldanha E. External Jugular Cutdown Technique for Totally Implantable Venous Access Devices: a Single-Centre Study. Indian J Surg Oncol. 2020;11(3):418-422. [Free full text]
- Bawazir OA, Altokhais TI. Hickman central venous catheters in children: open versus percutaneous technique. Ann Vasc Surg. 2020;68:209-216. [Available at]
- Sidoti A, Brogi E, Biancofiore G, et al. Ultrasound- versus landmark-guided subclavian vein catheterization: a prospective observational study from a tertiary referral hospital. Sci Rep. 2019;9(1):12248. [Free full text]
- Mansfield PF, Hohn DC, Fornage BD, et al. Complications and failures of subclavian-vein catheterization. N Engl J Med. 1994;331(26):1735-1738. [Free full text]
- Iwasaki T, Hayashi Y, Ohnishi Y, et al. Prospective analysis of percutaneous central venous catheterization in infants <4.0 kg undergoing cardiac surgery. Pediatr Cardiol. 2004;25(5):503-505 [Free full text]
- ASA physical status classification system. American Society of Anesthesiologists. Accessed January 2022. [Available at]
- Psallidas I, Helm EJ, Maskell NA, et al. Iatrogenic injury to the intercostal artery: aetiology, diagnosis and therapeutic intervention. Thorax. 2015;70(8):802-804. [Free full text]
- Davies HE, Merchant S, McGown A. A study of the complications of small bore 'Seldinger' intercostal chest drains. Respirology. 2008;13(4):603-607. [Available at]
- Havelock T, Teoh R, Laws D, Gleeson F; BTS Pleural Disease Guideline Group. Pleural procedures and thoracic ultrasound: British Thoracic Society Pleural Disease Guideline 2010. Thorax. 2010;65 Suppl 2:ii61-ii76. [Free full text]
- Carney M, Ravin CE. Intercostal artery laceration during thoracocentesis: increased risk in elderly patients. Chest. 1979;75(4):520-522. [Available at]
- Helm EJ, Rahman NM, Talakoub O, et al. Course and variation of the intercostal artery by CT scan. Chest. 2013;143(3):634-639. [Available at]
- Mossman D. Physician impairment: when should you report? Curr Psychiatry. 2011;10(9):67-71. [Free full text]