Veno-venous extracorporeal membrane oxygenation (VV-ECMO) is an extracorporeal life support strategy used in critically ill patients with potentially reversible severe respiratory failure (SRF) refractory to conventional ventilatory strategies (1 , 2). Most patients supported with VV-ECMO also require mechanical ventilation via an endotracheal tube. Some will require prolonged mechanical ventilation and/or have a significant sedation burden, and in this group, percutaneous dilatational tracheostomy (PDT) can be used to facilitate weaning from mechanical ventilation, reduce sedation burden, and to improve patient extracorporeal membrane oxygenation (ECMO) circuit interaction. Extracorporeal Life Support Organization guidelines recommend consideration of tracheostomy in ECMO patients (3).
PDT in general ICU patients is considered relatively safe with a very low frequency of major complications (4–6), even in patients with severe thrombocytopenia (7) or severe liver disease and refractory coagulopathy (8). However, robust data regarding the safety of this procedure in patients supported with VV-ECMO are not yet available. The only data at present are provided by a recent retrospective study by a group based at two German university hospitals (9), who concluded that PDT is a safe procedure when performed by experienced operators with careful optimization of the coagulation state.
ECMO circuit-blood interaction activates the coagulation cascade, with effects on fibrinolysis, thrombin formation, and platelet function (10 , 11). These effects can be further enhanced if ECMO circuit becomes dysfunctional (12). Hemostatic imbalance may be difficult to predict, with the coexistence of both thrombotic and hemorrhagic risks. PDT is an invasive procedure which might confer significant hemorrhagic risk in ECMO patients. The present study aims to provide more data in this field by investigating factors that may be associated with increased rate of complications in VV-ECMO patients undergoing PDT.
MATERIALS AND METHODS
This is a single-center retrospective observational study that investigated a patient cohort from the ICU in St Thomas’ Hospital (National SRF center, United Kingdom). Medical records were retrieved from the Phillips CareVue IntelliVue Clinical Information Portfolio system, investigating the period from 01/01/2010 to 31/12/2015. Patients were selected if they were supported with VV-ECMO and mechanical ventilation for SRF and required PDT. Collection and analysis of anonymized data were independently authorized and approved by the Guys & St Thomas’ Hospital Audit Committee (No 6068/25-01-2016). The need for individual informed consent was waived for this retrospective analysis of data collected prospectively for routine care, with no breach of privacy or anonymity.
All VV-ECMO patients had ultrasound-guided percutaneously placed ECMO cannulae (Medtronic, Minneapolis, MN) and either a Permanent Life Support or Heart-Lung Support circuit (Getinge, Rastatt, Germany) both with centrifugal pumps, Quadrox oxygenators (Getinge), and Bioline coated (Getinge) circuits. All extracorporeal circuit components were heparin coated.
PDT was performed according to clinical criteria and an individualized risk-benefit assessment performed by the ECMO consultant. Patients underwent a PDT as per a standard protocol with the Ciaglia Blue Rhino (CBR; Cook Critical Care, Bloomington, IL) single-step dilator technique (13) under video-bronchoscopic endotracheal visualization. Either a Portex blue line ultracuffed tracheostomy tube (Smiths Medical, Kent, UK) or a Tracoe Twist-plus tracheostomy tube (TRACOE medical GmbH, BVMed, Nieder-Olm, Germany) was used.
In our institution, the percutaneous dilation technique is the standard tracheostomy technique used in VV-ECMO patients, unless there is an absolute contraindication (Gross distortion of the neck anatomy, lack of neck landmark points due to obesity, neck soft-tissue infection) where open surgical technique is preferred.
The standard anticoagulation approach in our institution for patients on VV-ECMO is 50 ΙU/kg heparin bolus with cannulation, followed by a heparin infusion titrated to an activated partial thromboplastin time ratio (aPTTr) of 1.5–2. In patients with bleeding complications, heparin is reduced or stopped. In all patients, the periprocedure IV heparin protocol was individualized according to the ECMO consultant’s decision.
Hematologic and coagulation variables were managed periprocedure according to local standard ECMO Guidelines targets (aPTTr < 1.5, international normalized ratio < 1.5, hemoglobin > 70 g/L, platelets > 100 × 109 cells/L and fibrinogen > 1.5 g/L, respectively). Specialist hematology advice was sought as required according to ECMO consultant’s decision.
Baseline Clinical Variables
Baseline demographic and clinical characteristics in all patients supported with VV-ECMO prior to tracheostomy are presented in Supplemental Table 1 (Supplemental Digital Content 1, http://links.lww.com/CCM/E102).
Complications that occurred within the 48-hour period following PDT were deemed to be directly associated with the procedure.
Major complications were defined as mortality, cardiac arrest, loss of airway, tracheal or bronchial wall injury, pneumothorax (requiring intervention), major hemorrhage from stoma site or airway (requiring transfusion or surgical intervention), disseminated intravascular coagulation (DIC), and/or sepsis (associated with tracheal stoma infection). ECMO circuit dysfunction requiring oxygenator change was also considered a major complication as per a previous study (9).
Minor complications were defined as local hemorrhage arising from either stoma site or airway (self-limiting or treated successfully with local compression, topical vasoconstrictive agents, and/or electrocauterization), pneumomediastinum or pneumothorax not requiring drainage, local subcutaneous emphysema, and/or local stoma infections (not causing sepsis). Complications were also adjusted for operator’s grade (ICU consultant and ICU specialist trainee) and for different heparin infusion management prior to PDT.
Hematologic and coagulation markers and type and quantity of transfused blood products were collected prior to PDT, 24, and 48 hours following PDT. Outcome measures included total length of ICU stay and duration of ECMO support before and after PDT.
Before analysis, all continuous variables were tested by Kolmogorov-Smirnov test for normal distribution. Continuous variables were evaluated and found to be largely nonnormally distributed and are therefore presented as median values (interquartile range). Categorical variables are detailed as counts and/or percentages. The Mann-Whitney U test was applied for nonparametric analyses of continuous variables. Analysis of categorical variables was performed using the chi-square test. A binary logistic regression analysis was also used to calculate, for each variable, the odds ratio and 95% CI for major complications (excluding pneumothorax). Due to our limited sample size in relation to the number of candidate predicting variables, we performed a preselection by testing all variables using univariate analysis. All variables that have shown a p value of less than 0.1 were included in the multivariate model. A receiver operating characteristic (ROC) curves analysis was used for detecting cut-off values. A p value of less than 0.05 was considered to be statistically significant. Statistical package for the Social Science (SPSS Version 20.0; SPSS, Chicago, IL) was used to perform all analysis.
During the study period a total of 255 mechanically ventilated patients received VV-ECMO. Within this cohort, a total number of 65 patients (25%) underwent PDT. Detailed description of patients’ baseline characteristics is contained in Supplemental Table 1 (Supplemental Digital Content 1, http://links.lww.com/CCM/E102) and further hematologic, coagulation, and inflammatory variables in Table 1.
Complications Following PDT
All complications presented in patients with VV-ECMO that underwent PDT are reported in Table 2. No deaths occurred as a direct result of the procedure. Twenty-nine patients (45%) developed at least one complication, 10 of whom (15%) developed at least one major complication.
Detailed description of the cause of major complications in patients with VV-ECMO is illustrated in Table 3.
Minor complications rate is reported in Table 2. Within this group, bleeding was either self-limiting (79%) or was resolved by application of topical vasoconstrictive agents (16%) or electrocauterization (5%).
There were no baseline differences with regard to demographic and clinical characteristics or in hematologic, coagulation or inflammatory variables between patients with major complications (group B) and patients without major complications (group A). ECMO circuit function variables prior to PDT insertion in both patients groups are presented in Table 4. There was no difference in ECMO days prior to PDT for group B when compared with group A (7.5 [IQR, 6.7–12] vs 8.0 [IQR, 6–12]; p = 0.9] and no difference in ICU severity scores (Acute Physiology and Chronic Health Evaluation-II: 18 [IQR, 12–23] vs 18.5 [IQR, 17–22]; p = 0.4 and Sequential Organ Failure Assessment: 8 [IQR, 7–13] vs 10 [IQR, 6–13]; p = 0.4).
The duration of heparin interruption prior to PDT was not associated with major complications. The majority of patients had the heparin infusion stopped within 4–12 hours of performing the PDT (n = 43), and this group accounts for eight of 10 major complications. Eighteen patients who had not been receiving heparin for more than 12 hours accounted for only two cases of major complication. Two patients had stopped heparin infusion in less than 2 hours, and two more were receiving heparin infusion in low fixed rate (500 IU/hr); however, neither of these groups had major bleeding.
Analyzing major complications according to operator seniority for tracheostomy (ICU consultant vs ICU specialist trainee), we did not observe any significant difference ([8/47; 17%] vs [2/18; 11%]; p = 0.7, respectively). Similarly, no difference was found according to operators grade for bronchoscopy ([4/25; 16%] vs [6/40; 15%]; p = 0.6, respectively).
No difference in major complication rate was found between those patients who had a Portex tracheostomy tube (n = 32) comparing with those who had a Tracoe twist plus tracheostomy tube (n = 33), (18% vs 12.5%; p = 0.7).
Group B received a greater proportion of blood products periprocedure as described in Supplemental Table 2 (Supplemental Digital Content 2, http://links.lww.com/CCM/E103).
Patients remained in ICU for a median of 30 days (IQR, 19–44 d). Total ECMO supportive treatment was offered for a median of 16 days (IQR, 10–26 d). The overall ICU mortality outcome in our ECMO tracheostomized population was 28%. No difference was found in mortality at hospital discharge between patients with and without major complications (30% vs 27%, p = 0.86). Patients in group B had an increased total ECMO support days (26 [IQR, 17–39] vs 15 [IQR, 10–23]; p = 0.03) compared with patients in group A with no difference in ECMO support days prior to PDT.
Binary Logistic Regression Analysis
Among the variables tested in the univariate logistic regression analysis, ECMO postoxygenator PO2 and sweep gas flow were associated with major complications and included in the final multivariate logistic regression analysis. In this analysis, postoxygenator PO2 remained the only independent predictor of major complications periprocedure. A detailed description of this analysis is presented in Table 5.
ROC Curves Analyses
Further analyzing these data with ROC curves, we found that the area under the curve for ECMO postoxygenator PO2 was 0.764 ± 0.093; p value equals to 0.012; 95% CI, 0.58–0.95. The cut-off value of 40 kPa had 93% sensitivity and 56% specificity, and the cut-off value of 56 kPa had 56% sensitivity and 78% specificity.
We found that during our 6-year single-center experience, PDT in SRF patients supported with VV-ECMO carried a significant risk of major complications (15%) and far exceeds that previously reported in non-ECMO patients (14 , 15), where only 3% of patients suffered early severe complications. Major bleeding requiring blood transfusion or intervention was the most common major complication in our study, occurring in seven cases (11%), and accounted for approximately two thirds of all major complications. Mechanical complications occurred in five cases and included two posterior tracheal wall injuries, one bronchial mucosal injury, a thyroid isthmus, and external jugular vein injury, and one tension pneumothorax. ECMO circuit change was required in five cases (8%) due to either oxygenator failure or bleeding and circuit-related DIC. Major complications were associated with an increased blood product transfusion rate following completion of the PDT procedure, and these patients had prolonged ECMO support duration. Minor complication rates were also increased compared with previously reported general ICU studies (15–17). Most of the minor complications were self-limiting or treated conservatively with local hemostasis.
Periprocedure Management Of Coagulation
To our knowledge, there has been only one previous study that investigated the safety of performing PDT in ECMO patients (9). In that study, authors found PDT to be a relatively safe procedure with a low rate (4.2%) of clinically relevant major complications. The reasons for the higher complication rate in our study are not clear. Both studies applied similar PDT technique and patients had similar coagulation characteristics prior to PDT. The heparin management strategies were different between the two studies; the previous study followed a protocol that stopped heparin 1 hour before the planned tracheostomy and recommenced it immediately after without increased incidence of hemorrhage, whereas in our study, patients received an individualized heparin protocol according to clinician’s decision. Although there was no association between duration of heparin interruption prior to PDT and major complications in our cohort, the relatively small sample size does not allow for any definitive conclusions to be drawn, and further work is needed to define the optimal regimen.
The impact of ECMO on hemostasis is complex and includes thrombocytopenia, coagulation factor disturbance, hypofibrinogenemia, and acquired von Willebrand syndrome (AVWS) (10 , 11). Flow rates were numerically higher in the group that experienced complications (nonsignificant), and it may well be that the increased shear stress at higher flow rates induces AVWS (10), but further work is needed to explore this relationship further.
ECMO Membrane Dysfunction and Major Complications
An intriguing finding in our study is the lower preprocedure ECMO postoxygenator PO2 in the group of patients who experienced major complications, and this emerged as an independent predictor of major complications. This association has not been reported in previous work (9). Postoxygenator PO2 is a useful marker of ECMO membrane function, and a reduction in this value can indicate impending circuit failure, which is often due to microthrombus formation in the oxygenator membrane. Circuit-related coagulation disorders have been well documented (12), and it may well be that subclinical circuit-related DIC is “unmasked” by the PDT, and these patients then go on to develop bleeding and/or clinically apparent DIC.
Five patients needed ECMO circuit change post PDT due to primary ECMO oxygenator dysfunction or secondary to bleeding and DIC. Compared with the previous study (9), we performed PDT later (median 8 vs 3 d), with lower pump flow speeds (3.2 vs 3.9; L/min), and more prolonged heparin cessation. Together with the transfusion of blood products peri-PDT, this may have made vulnerable circuits more likely to fail (18) and increase our complication rate.
Given that ECMO oxygenator dysfunction prior to PDT emerged as a significant prognostic factor for major complications, further work should be undertaken to explore this association and whether optimizing ECMO circuit function prior to PDT might prevent complications.
The incidence of mechanical complications related to PDT in ECMO patients was higher than that we have experienced previously in non-ECMO patients. Of the five mechanical complications that occurred, two of these were thought to be relatively minor tracheal/bronchial mucosal “injury” (not rupture) due to the wire and/or bronchoscope itself. Despite this, four of the five mechanical complications resulted in major hemorrhage. ECMO patients often have very severe lung injury with friable tracheal and bronchial mucosal surfaces, and one explanation might be, therefore, that when such mucosal injury occurs, ECMO patients are more likely to have a significant bleed from this injury, that might otherwise have been clinically insignificant (or unrecognized), and a failing ECMO oxygenator increases the likelihood of this bleeding further.
Recent recommendations do not support the routine use of bronchoscopy during PDT in critically ill patients (19), and although it appeared not to have prevented relatively minor injury to either the tracheal wall or bronchial mucosa during PDT in our study, we would still recommend its use in the ECMO group for a number of reasons: As we have seen in our study, it is important to be able to identify when an injury has occurred and to have a low threshold for repeat bronchoscopy post PDT, particularly if bleeding is suspected. With this strategy bleeding can be identified early, specific therapies initiated, or additional expertise sought. With routine use of bronchoscopy for PDT, technical ability will improve for both senior and trainee intensivists, enabling more rapid identification and management of mechanical complications following PDT if and when they do occur, and the prevention of further problems during tracheostomy decannulation (20).
Pneumothorax was a rare event, presenting in only one of our patients. This is comparable with the 2% incidence reported in previous studies (21) and is within the range of the general ICU population (0–4%) (15 , 16 , 22).
Injury to thyroid and external jugular veins complicated with major hemorrhage occurred in one patient in our study. This was lower than that reported in a previous study (23), where 24 critically ill patients (5%) had bleeding due to vascular injury. Thyroid injury may be minimized through the use of ultrasound imaging techniques and in addition offers benefits with regard to optimal tracheal puncture site, with fewer adverse events and low incidence of minor bleeding (24).
Clinicians should recognize that PDT in ECMO patients carries a significant risk and take steps to minimize this risk where possible. This may include optimizing patient coagulation status, having a low threshold for changing the ECMO circuit prior to PDT if a failing membrane is suspected, and the routine use of bronchoscopy. PDT should be performed by experts familiar with both PDT procedure and ECMO patients. We recommend routine use of bronchoscopy during PDT to minimize bronchial/tracheal injury, identify and manage bleeding should it occur, and improve training and local expertise with its use. The optimal timing for PDT remains unclear, and there is currently a large variability in practice (25). It may well be that delaying PDT until the patient is liberated from ECMO is an acceptable approach. Further work is needed to define the optimal periprocedure heparin regimen. ECMO patients and/or next of kin should always be informed of the high risk PDT procedure.
Several potential biases limit the interpretation of these results, primarily those inherent in the retrospective study design. Minor complications of reduced clinical relevance may be poorly documented or omitted entirely, potentially causing underestimation. However, severe complications were likely well documented and thus detectable. Collection of follow-up data regarding long-term complications (such as tracheomalacia, tracheal wall abnormalities and oropharyngeal dysphagia) were missing mainly due to patient repatriation to referring hospitals. Despite the large number of clinical variables tested in the univariate regression analysis, other variables might have been accounted for major complications but not included in the analysis.
PDT is associated with a considerable complication rate in VV-ECMO patients. Impaired ECMO oxygenator function prior to PDT is an independent predictor of major complications and should always be a concern for clinicians prior to performing PDT in VV-ECMO patients.
We would like to thank St Thomas Hospital ICU Audit team for their significant contribution to patients’ data retrieval of our study
1. Allen S, Holena D, McCunn M, et al. A review of the fundamental principles and evidence base in the use of extracorporeal membrane oxygenation
(ECMO) in critically ill adult patients. J Intensive Care Med 2011; 26:13–26
2. Peek GJ, Mugford M, Tiruvoipati R, et al; CESAR Trial Collaboration: Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation
for severe adult respiratory failure (CESAR): A multicentre randomised controlled trial. Lancet 2009; 374:1351–1363
3. Brogan TV, Lequier L, Lorusso R, et al; Extracorporeal Life Support: The ELSO Red Book, Red Book 5th Edition. 2017Ann Arbor, Extracorporeal Life Support Organization.
4. Kost KM. Endoscopic percutaneous dilatational tracheotomy: A prospective evaluation of 500 consecutive cases. Laryngoscope 2005; 115:1–30
5. Terragni PP, Antonelli M, Fumagalli R, et al. Early vs late tracheotomy for prevention of pneumonia in mechanically ventilated adult ICU patients: A randomized controlled trial. JAMA 2010; 303:1483–1489
6. Young D, Harrison DA, Cuthbertson BH, et al; TracMan Collaborators: Effect of early vs late tracheostomy
placement on survival in patients receiving mechanical ventilation
: The TracMan randomized trial. JAMA 2013; 309:2121–2129
7. Kluge S, Meyer A, Kühnelt P, et al. Percutaneous tracheostomy
is safe in patients with severe thrombocytopenia. Chest 2004; 126:547–551
8. Auzinger G, O’Callaghan GP, Bernal W, et al. Percutaneous tracheostomy
in patients with severe liver disease and a high incidence of refractory coagulopathy: A prospective trial. Crit Care 2007; 11:R110
9. Braune S, Kienast S, Hadem J, et al. Safety
of percutaneous dilatational tracheostomy
in patients on extracorporeal lung support. Intensive Care Med 2013; 39:1792–1799
10. Heilmann C, Geisen U, Beyersdorf F, et al. Acquired von Willebrand syndrome in patients with extracorporeal life support (ECLS). Intensive Care Med 2012; 38:62–68
11. Malfertheiner MV, Philipp A, Lubnow M, et al. Hemostatic changes during extracorporeal membrane oxygenation
: A prospective randomized clinical trial comparing three different extracorporeal membrane oxygenation
systems. Crit Care Med 2016; 44:747–754
12. Lubnow M, Philipp A, Dornia C, et al. D-dimers as an early marker for oxygenator exchange in extracorporeal membrane oxygenation
. J Crit Care 2014; 29:473.e1–473.e5
13. Byhahn C, Wilke HJ, Halbig S, et al. Percutaneous tracheostomy
: Ciaglia blue rhino versus the basic ciaglia technique of percutaneous dilational tracheostomy
. Anesth Analg 2000; 91:882–886
14. Fernandez L, Norwood S, Roettger R, et al. Bedside percutaneous tracheostomy
with bronchoscopic guidance in critically ill patients. Arch Surg 1996; 131:129–132
15. Dempsey GA, Grant CA, Jones TM. Percutaneous tracheostomy
: A 6 yr prospective evaluation of the single tapered dilator technique. Br J Anaesth 2010; 105:782–788
16. Terragni PP, Antonelli M, Fumagalli R, et al. Early vs late tracheotomy for prevention of pneumonia in mechanically ventilated adult ICU patients: A randomized controlled trial. JAMA 2010; 303:1483–1489
17. Young D, Harrison DA, Cuthbertson BH, et al; TracMan Collaborators: Effect of early vs late tracheostomy
placement on survival in patients receiving mechanical ventilation
: The TracMan randomized trial. JAMA 2013; 309:2121–2129
18. Annich GM. Extracorporeal life support: The precarious balance of hemostasis. J Thromb Haemost 2015; 13(Suppl 1):S336–S342
19. Raimondi N, Vial MR, Calleja J, et al; FEPIMCTI and LACCTIN: Evidence-based guidelines for the use of tracheostomy
in critically ill patients. J Crit Care 2017; 38:304–318
20. Soni KD, Kumar A, Aggrawal R, et al. Do we need bronchoscopy during percutaneous tracheostomy
? J Anaesthesiol Clin Pharmacol 2016; 32:541–543
21. Kumar M. Subcutaneous emphysema and pneumothorax during percutaneous tracheostomy
without any evidence of tracheal wall injury on repeated bronchoscopy. Korean J Anesthesiol 2013; 65:589–590
22. Berg LF, Mafee MF, Campos M, et al. Mechanisms of pneumothorax following tracheal intubation. Ann Otol Rhinol Laryngol 1988; 97:500–505
23. Muhammad JK, Major E, Wood A, et al. Percutaneous dilatational tracheostomy
: Haemorrhagic complications and the vascular anatomy of the anterior neck. A review based on 497 cases. Int J Oral Maxillofac Surg 2000; 29:217–222
24. Alansari M, Alotair H, Al Aseri Z, et al. Use of ultrasound guidance to improve the safety
of percutaneous dilatational tracheostomy
: A literature review. Crit Care 2015; 19:229
25. Camporota L, Nicoletti E, Malafronte M, et al. International survey on the management of mechanical ventilation
during ECMO in adults with severe respiratory failure
. Minerva Anestesiol 2015; 81:1170–1183, 77 p following 1183