We have presented a patient with mediastinitis, pneumomediastinum, retropharyngeal and subcutaneous thoracic emphysema secondary to an un-anticipated PICC line extravasation. PICC lines, like all medical procedures carry risk. This is often under-appreciated and overlooked in complex patients.
When unanticipated clinical deterioration occurs in a patient, iatrogenic causes must be considered. In particular, the deployment of any invasive lines including a PICC line, must be carefully and diligently assessed as causative. In this case, complex multi-factorial presentation may have obscured a seemingly benign causative factor, such as a PICC line migration and vessel rupture.
The initial placement of the veno-venous bypass cannula was intra-thoracic, which resulted in acute haemothorax. However, with chest drain insertion and supportive management, this complication had resolved. Our patient was progressing well clinically until the development of infective signs and symptoms, accompanied by shortness of breath and voice changes on day 11. The temporal relationship is fitting, as the PICC line was inserted on day 10 and total parenteral nutrition started at the same time.
In our patient, the Computerised Tomography showed a PICC line tip of uncertain intra-luminal location on day11 post-operatively. It was beneficial to cease using the PICC line immediately following the radiological and clinical query on post-operative day 11, however no definitive decision as to its removal was made and the PICC line remained in situ. As illustrated in cases where the PICC line had extravasated into the mediastinum, followed by the infusion of high volume intravenous fluid, this can result in a pericardial tamponade with variable mortality reports. From the perspective of the risk of anatomical complications, it was beneficial that the tip of the PICC line was located well outside the pericardial sac, as recommended by a number of authors. The resulting complications of extravasation were therefore less likely to cause a pericardial tamponade.
Pneumo-mediastinum may occur spontaneously or can be secondary, with numerous causative factors such as lung disease, barotrauma, trauma or iatrogenic causes (6). It is recognised through laboured breathing and shortness of breath and can be associated with voice change, retro-pharyngeal air tracking and subcutaneous emphysema. Our patient exhibited all of these symptoms, which resolved gradually with appropriate intraoperative removal of aetiological cause and supportive care in the post-operative period. It is likely that the extravasation of total parenteral nutrition caused the inflammatory changes and creation of air pockets. This pathophysiological pattern has been reported previously by a group of authors, however the severity and anatomical air distribution differed in our case, with greater retro-pharyngeal tracking and potential airway compromise (7).
Mediastinitis can be a challenging diagnosis to deduce as it presents with a non-specific confluence of signs and symptoms. It is often a diagnosis of exclusion with a wide and varied differential diagnosis. Signs and symptoms can consist of chest pain, dyspnea, hypoxemia, neck swelling, difficulty swallowing, dilated superficial and neck veins. Differential diagnosis of mediastinitis consists of number of alternative high morbidity conditions such as aortic dissection, pneumonia, pleural effusions, mediastinal masses or abscess, pericardial effusion or tamponade. The non-specificity of clinical signs and symptoms may lead to a delayed diagnosis or misdiagnosis, which contributes to patients’ morbidity and mortality (8).
Recommended site of placement of tip of a central venous catheter is above the level of the right atrium in the lower or upper superior vena cava, as outlined in guidelines by the United States Food and Drug Administration (FDA) (9-10). In clinical terms, the anatomical level corresponds to the level of the carina, as delineated in anatomical cadaveric and radiographic studies (10). Pericardium crosses the SVC below the level of the carina, with the pericardial sac therefore being below this point (10). This end point of pericardium has been outlined in both anatomical and cadaveric studies (3) (10). By ensuring that the tip of the PICC is above the level of the carina, the risk of pericardial tamponade due to vessel rupture can be minimised. A small number of authors recommend that the peripherally inserted central venous catheters should be located at the junction of superior vena cava and right atrium. This is controversial as it may lead to cardiac rupture and tamponade (9). It is recommended for the tip not to abut the vein or the heart wall at an acute angle, as incidence of perforation increases as the degree of perpendicular orientation of the catheter with the wall of the superior vena cava increases (10)(11).
In our patient, the presence of the PICC line was confirmed with a plain chest radiograph on the day of the insertion to be in the left sided subclavian vein. This is not considered an ideal location by some authors including the USA FDA. Other authorities however consider as large a vessel as possible outside the pericardial sac typically the superior vena cava as acceptable location (12). Subclavian vein location, as opposed to a larger vessel such as superior vena cava, may have been a contributing factor to the extravasation due to vessel erosion by the hypertonic total parenteral nutrition solution.
Peripherally inserted Central Venous catheters are known for their potential to migrate with arm movement. In a cadaver study, it was demonstrated that the antecubital PICC lines can advance up to 7 cm with abduction and elevation of the arm (13). An in vivo study found a lower but significant degree of movement with 58% of PICC lines advancing 2 cm or more with arm abduction (14). Haygood et conducted a retrospective review of central lines which had migrated into the azygous vein. Four of these catheters were PICC lines, all of which were inserted through the left side (15). Left sided catheters may also be more likely to result in vessel rupture. When a left sided catheter is not inserted far enough, the catheter tip is more likely to be positioned at a perpendicular angle to the vessel wall, resulting in a compromise of structural wall integrity (16)
PICC’s are being increasingly used in critically ill patients. The insertion site is perceived to be of lower risk as it minimises the risk of pneumothorax, haemothorax and can be used with greater degree of safety in coagulopathic patients. Some authors suggest that PICC lines may be more favourable in patients with severe cardiorespiratory abnormalities and morbid obesity (17). Challenging the clinical perception that peripherally inserted central venous catheters may be safer as compared to central venous lines, a recent review found that malpositoning of the catheter tip, thrombophlebitis and catheter dysfunction were more common than with central venous lines (18). In a meta-analysis of 11 studies comparing the risk of deep vein thrombosis related to PICCs with that related to CVCs it was shown that PICCs were associated with an increased risk of deep vein thrombosis (OR 2.55, 1.54–4.23, p<0.0001) but not pulmonary embolism (no events) (4). The association was particularly strong in critically ill patients and those with a malignancy.
In a prospective cohort study of PICC catheters in neonates, a 1.5% incidence of cardiac tamponade was found among the 194 patients who had this type of catheter inserted routinely. The authors concluded that the association of pericardial effusion with a PIC catheter is a rare but extremely serious event, requiring the use of bedside echocardiography (19). To our knowledge there has been no known, equivalent study performed in adult population.
Cardiac tamponade is more frequently associated to central venous catheters inserted through a peripheral vein, rather than those placed directly in a central vein (20). There are other rare serious complications associated with PICC lines such as catheter fracture and arrhythmias. This is attributed to the mobility and migration of peripherally inserted central venous catheters. There have been literature reports of pericardial tamponade due to the presumed migration of a peripherally inserted central venous catheter (21). In a 20 year-old patient, the tip of the PICC line was identified within the right atrium, migrated and resulted in cardiovascular arrest due to the extravasation of potassium into the pericardial fluid from the PICC. In a 14 year-old patient, where a PICC insertion was approached from the left ante-cubital fossa, cannulation of the right pericardiophrenic vein led to the development of cardiac tamponade (22).
We have reported on delayed migration of a peripherally inserted central venous catheter into the chest cavity as a causative factor for extensive retropharyngeal and subcutaneous emphysema, pneumo-mediastinum, as well as mediastinitis in a patient with a recent dual liver and renal transplant.
Although there is a perception of relative safety of PICC lines in critical care patients, this is being challenged by recent systematic reviews demonstrating increased risk of deep venous thrombosis compared to central venous catheters. In addition, there is a high frequency of PICC line malposition and migration, which is associated with malfunction and potential vessel rupture. Although rare, vessel rupture can result in mediastinal complications with associated morbidity and mortality if unrecognized. Migration, misplacement and potential catheter pericardial migration need to be considered in the differential diagnosis of intra-thoracic pathology including pneumo-mediastinum with retropharyngeal air tracking, subcutaneous emphysema, mediastinitis, pleural complications and pericardial tamponade.
This clinical and radiological enquiry should be performed both as part of a differential diagnosis and prior to PICC line use in critically ill patients.