Arterial Complications Assessed by Duplex Ultrasound After Decannulation of Peripheral Venoarterial Extracorporeal Membrane Oxygenation
Article ID: CJ-24-0400
Details
Article ID: CJ-24-0400
Background: Vascular complications are common and can be fatal even after successful decannulation in patients with peripherally cannulated veno-arterial extracorporeal membrane oxygenation (VA-ECMO). Therefore, we aimed to accurately determine the incidence of arterial complications assessed by Duplex ultrasound following peripheral VA-ECMO decannulation. In addition, we investigated the predictors of severe complications requiring intervention.
Methods and Results: We retrospectively reviewed 1,350 adult patients who underwent ECMO between January 2012 and April 2023. Of 839 patients treated with peripherally cannulated VA-ECMO, 596 were successfully weaned off and 212 underwent Duplex ultrasound for final analysis. The primary outcome was arterial complications requiring vascular intervention. Thirty-three (15.6%) patients experienced such complications after decannulation. Acute limb ischemia due to thrombotic occlusion was the most common complication, occurring in 23 (10.8%) patients, followed by stenosis (3.8%), pseudoaneurysm (3.8%), arteriovenous fistula (0.9%), and dissection (0.9%). No significant differences in complication rates were found between the percutaneous and surgical decannulation groups in the propensity score-matched population (12.7% vs. 15.9%, respectively; P=0.799). Multivariable analysis revealed disseminated intravascular coagulation (DIC; odds ratio 2.6; 95% confidence interval 1.17–5.69; P=0.019) as the only predictor of arterial complications after decannulation.
Conclusions: Arterial complications requiring vascular intervention frequently occur following successful weaning from VA-ECMO regardless of the decannulation strategy. In this setting, DIC appears to be associated with an increased rate of arterial complications.
Veno-arterial extracorporeal membrane oxygenation (VA-ECMO) is indicated in refractory cardiogenic shock, where low cardiac output makes it difficult to maintain adequate tissue perfusion to meet oxygen demand despite medication therapy.1 In recent years, advances in extracorporeal membrane oxygenation (ECMO) equipment and shock management have led to a marked increase in the use of VA-ECMO worldwide.2 Nevertheless, the mortality rate of patients on VA-ECMO remains high, and efforts to reduce ECMO-related complications as much as possible are crucial, especially because these complications can be associated with mortality.3
Vascular complications associated with ECMO cannulation and maintenance are one of the major complications of ECMO, especially because peripheral VA-ECMO typically involves a femoral artery puncture, resulting in arterial complications at the puncture site.4 Duplex ultrasound is a very effective modality for identifying vascular complications of femoral artery puncture.5 A vascular closure device (VCD) was first introduced to reduce femoral arterial complications in coronary interventions and improve quality of care.6 Surgical decannulation has been the traditional method for ECMO weaning. Since we first reported in 2016 that VCD can be a safe and effective method for decannulation of VA-ECMO, several studies have been published on the use of VCD in VA-ECMO decannulation.7–9 However, most previous studies on ECMO-related vascular complications have small sample sizes and studies that systematically evaluated vascular complications using Duplex ultrasound after decannulation are lacking. Therefore, the primary purpose of the present study was to accurately determine the incidence of arterial complications requiring intervention after decannulation, as assessed by Duplex ultrasound, in patients successfully weaned from peripheral VA-ECMO using data for 10-year period from a single-center registry. In addition, we aimed to analyze the predictors of these complications.
We retrospectively analyzed data for 1,350 consecutive adult patients in whom ECMO was effectively initiated between January 2012 and April 2023 from the Samsung Medical Center ECMO Registry in the Republic of Korea (Figure 1). We identified 839 patients treated with peripheral VA-ECMO after excluding 327 patients in whom venovenous ECMO was initiated, 133 patients in whom central ECMO was initiated, and 51 patients in whom the ECMO mode or configuration was changed. Of these 839 patients, 243 either died while on ECMO support or were weaned from ECMO under conditions deemed medically futile. Ultimately, 596 patients were successfully weaned from ECMO and, of these, 212 underwent Duplex ultrasound after decannulation. Our analysis centered on the characteristics, ultrasound findings, and clinical outcomes of these 212 patients.
Flowchart showing patient recruitment to this study from the Extracorporeal Membrane Oxygenation (ECMO) Registry.
The Institutional Review Board (IRB) of Samsung Medical Center approved this study (IRB no. 2020-10-102) and waived the requirement for informed consent because of the observational nature of the study. Patient information was anonymized and deidentified before analysis.
ECMO CannulationFinal decisions on ECMO initiation and the cannulation method were made jointly by ECMO experts, including cardiac surgeons, interventional cardiologists, and critical care physicians. All cannulations were performed by experienced operators. ECMO cannulation and management approaches at Samsung Medical Center have been described previously.10,11 Most cannulations were performed percutaneously using the Seldinger technique with ultrasound guidance or femoral artery pulse palpation. In a few cases where percutaneous cannulation failed, surgical cannulation was achieved by exposing the vessel using a cut-down technique. In situations where fluoroscopy was available, fluoroscopy-guided cannulation was performed, and bedside insertion was performed in the remaining cases. The oxygenator was the Capiox Emergency Bypass System (Capiox EBSTM; Terumo, Inc., Tokyo, Japan) and Permanent Life Support (PLS; MAQUET, Rastatt, Germany). The diameter of the return cannula inserted into the femoral artery varied between 14 and 21 Fr, whereas the diameter of the drain cannula inserted into the femoral vein ranged from 20 to 28 Fr. If possible, a catheter for preventing distal limb ischemia was preemptively inserted distal to the arterial cannulation site, and, if not, inserted later upon occurrence of distal limb ischemia. The adequacy of the cannulation position was checked immediately after ECMO cannulation using X-ray imaging. Unfractionated heparin was chosen as the initial anticoagulant unless there was active bleeding. The heparin infusion rate was adjusted through Samsung Medical Center’s protocol, with target activated clotting times of 150–180 s and activated partial thromboplastin times of 55–75 s. After ECMO initiation, early left ventricular decompression was performed by percutaneous atrial septostomy or surgical venting in the absence of aortic valve opening, low pulse pressure, and worsening pulmonary edema despite diuretics, continuous renal replacement therapy, or inotropes.
ECMO DecannulationAt Samsung Medical Center, vasodilator and inotrope requirements were norepinephrine ≤0.05 µg/kg/min and dobutamine ≤5 µg/kg/min, mean arterial pressure maintained at at least 65 mmHg after 1 h of ECMO flow of at least 1 L/min, cardiac output measured by echocardiography of at least 3 L/min, tissue Doppler imaging of the right ventricular free wall at least 9 cm/s, central venous pressure at or below 15 mmHg, or, if the clinician deemed it feasible, we considered weaning from VA-ECMO. Following a definitive decision to remove VA-ECMO support, anticoagulant administration was halted several hours before commencing the decannulation procedure.
For percutaneous ECMO decannulation, we used a post-closure technique with 2 Perclose ProGlides (Abbott Vascular, Clonmel, Ireland). The methodology for VCD-based decannulation at Samsung Medical Center has been described previously.7 We directly punctured the arterial cannula at its proximal end and threaded a 0.035-inch guidewire (Curved Terumo Wire; Terumo, Tokyo, Japan) through it. Upon removal of the arterial cannula, we positioned the first ProGlide device into the femoral artery using the guidewire. We then removed the guidewire and applied the first suture. Upon reinserting the guidewire through the first ProGlide’s hole, we removed this device and deployed the second ProGlide into the femoral artery, after which the second suture was applied. Subsequently, the second ProGlide was removed by retracting the guidewire through its hole. An assistant continually applied manual pressure to secure the guidewire within the artery. Knots were fastened with a knot pusher, and we verified adequate hemostasis before removing the guidewire and applying final pressure to the site.
Our surgical protocol entailed an initial cut-down incision at the cannulation site to expose the femoral artery. The arterial cannula was then withdrawn after the purse-string suture was secured. In cases of severe arterial damage, vascular repair was executed via a bovine pericardial patch at the surgeon’s discretion. If a thrombus was detected or blood backflow was deemed insufficient, a Fogarty balloon catheter was used for thrombectomy. Finally, the surgical site was closed using sutures.
The approach to removal of the distal perfusion catheter was dictated by the method used for arterial decannulation. In instances of percutaneous decannulation, the distal perfusion catheter was extracted with subsequent manual compression to the site. Conversely, if surgical decannulation was performed, the distal perfusion catheter was also removed surgically.
Duplex Ultrasound and Arterial ComplicationsFollowing ECMO decannulation, Duplex ultrasound was performed once on the day of decannulation to evaluate both vascular integrity and blood flow, spanning from the common iliac artery to the below-the-knee arteries in the lower extremity where the perfusion cannula had been inserted. Complications were defined as abnormalities detected via Duplex ultrasound, and the severity of complications was determined based on whether they required intervention. Only vascular complications that required either surgical or interventional management were considered clinically significant in our study. The overall complications comprised hematoma, thrombotic occlusion, arterial stenosis, pseudoaneurysm, arteriovenous fistula, arterial dissection, and infection. All Duplex ultrasound imaging was performed by well-trained and licensed vascular technicians. We also evaluated patients for disseminated intravascular coagulation (DIC) to identify risk factors for arterial complications. DIC was defined as a DIC score of ≥5 on the scoring system developed by the International Society on Thrombosis and Hemostasis.12 The DIC score was calculated twice. First, it was based on laboratory results obtained within 24 h before ECMO insertion and within 48 h after insertion, referred to as the “cannulation DIC score”. Second, it was calculated using the results obtained within 48 h before and after ECMO decannulation, with the value closest to the time of decannulation used for the “decannulation DIC score”.
Statistical AnalysisTo compare characteristics and clinical outcomes between the 2 groups, we used Chi-squared or Fisher’s exact tests for categorical variables, when applicable, and the Mann-Whitney U test for continuous variables. Categorical data are presented as numbers and percentages, and continuous variables are presented as the mean±SD or as the median with interquartile ranges. Variables with P<0.05 on univariate analyses and clinically relevant variables were entered into the multiple logistic regression model to investigate predictors for complications requiring intervention and in-hospital mortality. Results are reported as odds ratios (ORs) and 95% confidence intervals (CIs) for each variable. All analyses were performed using R version 4.3.1 (R Foundation for Statistical Computing, Vienna, Austria).
Patients’ baseline characteristics are presented in Table 1. There were no significant differences in comorbidities and laboratory findings between patients with and without complications requiring intervention. When comparing laboratory results at the time of ECMO insertion, there were no significant differences in platelet count, prothrombin time, D-dimer, and fibrinogen levels between the 2 groups. However, the incidence of DIC, defined as a score of ≥5, was significantly higher in the group with than without complications. In addition, when comparing laboratory results at the time of ECMO decannulation, the platelet count was significantly lower and the DIC score was significantly higher in the group with than without complications. There were no significant differences between the 2 groups in terms of antiplatelet therapy, including the use of single or dual antiplatelet therapy. Factors linked to ECMO, such as extracorporeal cardiopulmonary resuscitation, fluoroscopy-guided cannulation, cannula size, and whether both arterial and venous cannulas were placed in one groin, did not differ significantly between the groups with and without complications.
Baseline Characteristics of Patients With Successful ECMO Decannulation According to the Presence or Absence of Complications Requiring Intervention
| No intervention (n=179) |
Requiring intervention (n=33) |
P value | |
|---|---|---|---|
| Age (years) | 58.1±16.3 | 61.4±14.6 | 0.280 |
| Male sex | 114 (63.7) | 16 (48.5) | 0.146 |
| BMI (kg/m2) | 24.4±4.3 (179) | 23.7±3.7 (32) | 0.395 |
| Current smoker | 38/178 (21.3) | 7/33 (21.2) | >0.999 |
| Medical history | |||
| Hypertension | 77 (43.0) | 20 (60.6) | 0.094 |
| Diabetes | 65 (36.3) | 18 (54.5) | 0.075 |
| Dyslipidemia | 37 (20.7) | 7 (21.2) | >0.999 |
| Malignancy | 21 (11.7) | 4 (12.1) | >0.999 |
| Chronic kidney disease | 28 (15.6) | 4 (12.1) | >0.799 |
| Peripheral artery disease | 7 (3.9) | 3 (9.1) | 0.399 |
| Stroke | 19 (10.6) | 3 (9.1) | >0.999 |
| Myocardial infarction | 43 (24.0) | 6 (18.2) | 0.612 |
| Patient status at ECMO initiation | |||
| SBP (mmHg) | 58.4±50.0 (133) | 53.3±41.5 (27) | 0.618 |
| DBP (mmHg) | 37.3±30.7 (133) | 32.6±25.0 (27) | 0.464 |
| Hemoglobin (g/dL) | 11.4±2.6 (110) | 11.1±2.5 (23) | 0.554 |
| Platelet count (×103/μL) | 199.3±87.5 (110) | 181.2±97.1 (23) | 0.377 |
| Prothrombin time (INR) | 2.3±1.9 (147) | 2.1±0.9 (27) | 0.555 |
| Fibrinogen (mg/dL) | 243.7±123.0 (134) | 229.8±123.9 (26) | 0.598 |
| D-dimer (μg/mL) | 11.5±14.7 (125) | 12.2±12.8 (24) | 0.833 |
| Total bilirubin (mg/dL) | 1.6±2.1 (98) | 1.4±1.8 (21) | 0.696 |
| Creatinine (mg/dL) | 2.0±2.3 (108) | 1.8±1.5 (21) | 0.617 |
| Lactic acid (mmol/L) | 5.9±4.7 (90) | 7.2±5.3 (20) | 0.278 |
| DIC score | 3.3±2.4 | 3.9±2.3 | 0.189 |
| DIC | 60 (33.5) | 18 (54.5) | 0.035 |
| Patient status at ECMO decannulation | |||
| Platelet count (×103/μL) | 106.5±100.6 | 82.5±52.9 | 0.047 |
| Prothrombin time (INR) | 1.3±0.8 | 1.3±0.3 | 0.928 |
| Fibrinogen (mg/dL) | 372.0±163.9 (166) | 366.7±156.0 (33) | 0.866 |
| D-dimer (μg/mL) | 11.8±15.0 (139) | 8.4±6.8 (31) | 0.064 |
| APTT (s) | 66.8±27.8 (178) | 65.0±20.4 (33) | 0.670 |
| DIC score | 3.1±1.5 | 4.1±1.2 | 0.001 |
| DIC | 29 (16.2) | 10 (30.3) | 0.094 |
| Antiplatelet therapy | 96 (53.6) | 15 (45.5) | 0.500 |
| Single antiplatelet therapy | 16 (8.9) | 3 (9.1) | >0.999 |
| Dual antiplatelet therapy | 80 (44.7) | 12 (36.4) | 0.486 |
| Parameters related to ECMO | |||
| ECPR | 76 (42.7) | 11 (33.3) | 0.417 |
| Fluoroscopy-guided cannulation | 73 (40.8) | 11 (33.3) | 0.542 |
| Arterial cannula size (Fr) | 15.0 [15.0–16.5] | 15.0 [15.0–15.0] | 0.201 |
| Venous cannula size (Fr) | 21.0 [21.0–22.0] | 21.0 [21.0–22.0] | 0.421 |
| Same groin cannulation | 112 (62.6) | 23 (69.7) | 0.558 |
| ECMO decannulation strategy | 0.038 | ||
| Manual compression | 1 (0.6) | 2 (6.1) | |
| Surgical decannulation | 87 (48.6) | 13 (39.4) | |
| Percutaneous decannulation | 91 (50.8) | 18 (54.5) | |
Unless indicated otherwise, data are presented as the mean±SD (n), median [interquartile range], or n/N (%). APTT, activated partial thromboplastin time; BMI, body mass index; DBP, diastolic blood pressure; DIC, disseminated intravascular coagulation; ECMO, extracorporeal membrane oxygenation; ECPR, extracorporeal cardiopulmonary resuscitation; INR, international normalized ratio; SBP, systolic blood pressure.
Arterial Complications Requiring Intervention
Of the 212 patients who underwent post-decannulation vascular assessment via Duplex ultrasound, 33 experienced arterial complications requiring intervention, yielding a complication rate of 15.6% (Figure 2). The predominant issue was acute limb ischemia due to thrombotic occlusion, observed in 23 (10.8%) patients and managed with thromboembolectomy or endarterectomy. Arterial stenosis and pseudoaneurysm were the next most frequent complications, affecting 8 (3.8%) patients each. Therapeutic interventions for arterial stenosis included stent implantation with balloon angioplasty or patch angioplasty. Of the 8 patients with confirmed pseudoaneurysms, 5 were treated with open aneurysm repair, 2 received thrombin injections, and 1 was managed with ultrasound-guided compression. Procedures for arteriovenous fistula and arterial dissection were performed in 2 (0.9%) patients each. Arterial vascular complications requiring intervention were significantly higher in cases of decannulation with manual compression than in cases of surgical or percutaneous decannulation (6.1% vs. 0.6%; P=0.038). However, there was no significant difference in the frequency of arterial complications requiring intervention between the percutaneous and surgical decannulation groups (16.5% vs. 13.0%, respectively; P=0.604; Figure 3; Supplementary Table 1).
Arterial vascular complications requiring intervention after decannulation. AVF, arteriovenous fistula.
Arterial vascular complications requiring intervention according to decannulation strategy: (A) percutaneous decannulation; (B) surgical decannulation. AVF, arteriovenous fistula.
Due to the substantial differences in baseline characteristics between the percutaneous and surgical decannulation groups, propensity score matching was performed to minimize these differences (Supplementary Table 2). Although the matching process improved comparability, it did not fully eliminate the baseline differences. Nevertheless, there remained no significant difference in the incidence of arterial complications requiring intervention between the percutaneous and surgical decannulation groups in the propensity score-matched population (12.7% vs. 15.9%, respectively; P=0.799; Supplementary Table 3).
In-Hospital Management and Clinical OutcomesThere were no significant differences in in-hospital management strategies between patients with and without complications requiring intervention (Table 2). There was no difference in the use of a distal perfusion catheter to prevent distal limb ischemia between the 2 groups. The duration of ECMO support also did not vary significantly, averaging 5.8 days in the group with complications and 6.9 days in the group without complications (P=0.443). There were no significant differences in other complications, including cannulation site bleeding, gastrointestinal bleeding, rhabdomyolysis, sepsis, and stroke, between the 2 groups. In-hospital mortality rates were similar between the groups with and without complications (24.2% vs. 14.0%, respectively; P=0.217).
In-Hospital Management and Clinical Outcomes in Patients With and Without Complications Requiring Intervention
| No intervention (n=179) |
Requiring intervention (n=33) |
P value | |
|---|---|---|---|
| In-hospital management during ECMO | |||
| Left heart decompression with venting | 28 (15.6) | 7 (21.2) | 0.591 |
| Distal perfusion | 124 (69.3) | 24 (72.7) | 0.849 |
| Continuous renal replacement therapy | 80 (44.7) | 14 (42.4) | 0.960 |
| Mechanical ventilation | 145 (81.0) | 26 (78.8) | 0.955 |
| Intra-aortic balloon pump | 8 (4.5) | 0 (0.0) | 0.459 |
| Vasopressor | 155 (86.6) | 30 (90.9) | 0.690 |
| Anticoagulation | 136/151 (90.1) | 26/28 (92.9) | 0.911 |
| Clinical outcomes | |||
| ECMO insertion site bleeding | 20/178 (11.2) | 6/33 (18.2) | 0.408 |
| Rhabdomyolysis | 9/178 (5.1) | 0/33 (0.0) | 0.395 |
| Stroke | 7/178 (3.9) | 2/33 (6.1) | 0.931 |
| Sepsis | 3/178 (1.7) | 0/33 (0.0) | >0.999 |
| Gastrointestinal bleeding | 8/178 (4.5) | 2/33 (6.1) | >0.999 |
| ECMO support duration (days) | 6.9±10.9 | 5.8±6.7 | 0.443 |
| In-hospital mortality | 25 (14.0) | 8 (24.2) | 0.217 |
Unless indicated otherwise, data are presented as the mean±SD (n) or n/N (%). Abbreviations as in Table 1.
Factors Associated With Complications Requiring Intervention
In multivariable logistic regression analysis, the DIC score at the time of ECMO decannulation was the only independent predictor of arterial complications, with an OR of 1.5 (95% CI 1.16–2.04; P=0.003). Age, extracorporeal cardiopulmonary resuscitation, cannulation strategy, decannulation method, and a medical history of hypertension, diabetes, or chronic kidney disease were not predictors of the occurrence of complications requiring intervention (Table 3).
Factors Associated With Vascular Complications Requiring Intervention
| OR (95% CI) | P value | |
|---|---|---|
| Age | 1.0 (0.97–1.03) | 0.875 |
| Hypertension | 2.4 (0.93–6.28) | 0.069 |
| Diabetes | 2.1 (0.86–5.04) | 0.106 |
| Chronic kidney disease | 0.4 (0.12–1.52) | 0.189 |
| ECPR | 0.6 (0.27–1.50) | 0.297 |
| DIC scoreA | 1.5 (1.16–2.04) | 0.003 |
| Fluoroscopy-guided cannulation | 0.6 (0.26–1.47) | 0.274 |
| ECMO support duration | 1.0 (0.95–1.05) | 0.842 |
| Percutaneous decannulation | 1.2 (0.54–2.80) | 0.622 |
ADIC score is based on the laboratory results closest to the time of ECMO decannulation. CI, confidence interval; OR, odds ratio. Other abbreviations as in Table 1.
Similarly, a medical history of hypertension, diabetes, or chronic kidney disease and other ECMO-related complications were not associated with in-hospital mortality (Supplementary Table 4). Acute limb ischemia due to complete thrombotic occlusion was significantly associated with an increase in in-hospital mortality, with an OR of 4.2 (95% CI 1.41–12.76; P=0.010).
In the present study, we investigated the incidence and predictors of arterial vascular complications requiring endovascular intervention or surgery after ECMO decannulation assessed by Duplex ultrasound. The major findings are as follows: (1) clinically relevant arterial complications occurred in 15.6% of patients successfully weaned from femorally cannulated VA-ECMO; (2) there was no significant difference in the incidence of arterial complications between surgical and percutaneous decannulation strategies; (3) the DIC score at the time of ECMO decannulation was the only predictive factor for arterial vascular complications; and (4) acute limb ischemia induced by thrombotic occlusion was significantly associated with increased in-hospital mortality. A key strength of our study lies in the use of Duplex ultrasound, enhancing the diagnostic reliability of vascular complications.
In previous studies, the rate of vascular complications during ECMO varied from 17% to 50%.3,4,13–23 Vascular complications in ECMO patients can be caused by various factors throughout ECMO management, such as at the time of cannulation, during ECMO maintenance, and after decannulation.4 In studies reporting vascular complications occurring after decannulation, rates have varied from 3% to 44%.7,8,24–30 However, most of these studies had small sample sizes, focused on decannulation strategies, and lacked analysis of the various factors that influence vascular complications. In addition, they reported the incidence based on acute complications with critical symptoms and signs that occurred immediately after decannulation, which may underestimate the true incidence of complications. Thus, we investigated the true incidence of vascular complications after decannulation via Duplex ultrasound, and arterial complications requiring intervention seemed to be high even after successful decannulation. The usefulness of Duplex ultrasound as a tool to detect vascular complications has already been confirmed in previous studies.5,31–35 Because we confirmed the vascular status after decannulation with Duplex ultrasound in addition to clinical symptoms, we were able to determine the incidence of vascular complications more precisely.
Previous studies have reported a wide range of risk factors for vascular complications related to ECMO. Among the many reported risk factors, younger age, female sex, history of peripheral arterial occlusive disease, the absence of a distal perfusion catheter, and surgical decannulation have been repeatedly reported as risk factors.3,4,8,13,17–19,21,26,29,36–38 In our study, the occurrence of arterial complications after decannulation was not significantly affected by the risk factors previously reported in the literature. When comparing percutaneous to surgical decannulation, excluding decannulation with manual compression, no significant difference in complication rates was found between the 2 groups. This study was a retrospective analysis, and there were substantial differences in baseline characteristics between the 2 groups. To minimize these differences, we performed propensity score matching. However, due to the relatively small sample size, it was challenging to fully eliminate the baseline differences. Nevertheless, even after reducing the differences between the groups as much as possible through propensity score matching, no significant difference in arterial complication rates was observed according to the decannulation strategy. All 3 cases of decannulation with manual compression in this study were incidental, rendering any meaningful analysis of the impact of manual compression on vascular complications unfeasible.
In our study, the only factor that influenced the occurrence of vascular complications was DIC. Severe inflammation due to severe shock or vascular injury can cause DIC, which not only increases the risk of bleeding but also increases the risk of thrombosis. DIC-induced thrombosis may have contributed to the acute limb ischemia that accounted for a large proportion of the vascular complications we reported. In addition, the occurrence of thrombocytopenia or prolongation of prothrombin time, which constitute the DIC score, may have contributed to the difficulty in adequately maintaining the anticoagulation therapy required in ECMO patients. Moreover, DIC is often observed in clinical settings involving severe conditions like sepsis, trauma, and shock. According to our data, the severity of the patient’s condition appears to be a more influential determinant of vascular complications following decannulation than any aspects directly associated with ECMO devices or procedures. These findings are consistent with earlier studies that demonstrated a correlation between elevated Sequential Organ Failure Assessment (SOFA) scores at the time of ECMO cannulation and a heightened risk of acute limb ischemia.8 In our study, acute limb ischemia resulting from thrombotic occlusion stood out as a significant factor contributing to elevated in-hospital mortality. In addition, in our separate multivariable analysis focused on complications such as thrombotic occlusion, the DIC score remained the only independent predictor of thrombotic occlusion. This highlights the importance of DIC as a key factor in vascular complications after decannulation. These results further reinforce the role of DIC in the development of vascular complications, particularly thrombotic events. Therefore, extra caution is recommended to prevent acute limb ischemia, particularly in patients removed from ECMO due to clinical improvement.
This study has several limitations. First, as a single-center retrospective study, this study has inherent limitations, including the possibility of selection bias and the influence of confounding factors. However, our research comprises the largest sample size in studies focusing on post-decannulation complications thus far, which we believe adds significant strength to our findings despite these limitations. The decannulation strategies, whether surgical or percutaneous, were determined by the attending physician, resulting in divergent baseline characteristics for the 2 groups. This variability constrains our ability to determine the effect of decannulation methodology on complication rates. We attempted to control for confounders through multivariable analysis, but residual confounding remains a concern. We performed a multivariable analysis using the DIC score at the time of ECMO decannulation and found that the DIC score was the only independent predictor of complications. However, even when using the DIC at the time of ECMO initiation as a variable, it still significantly predicted the occurrence of complications (Supplementary Table 5). This finding may be explained by the fact that the average duration of ECMO support in our cohort was <1 week, which suggests that the coagulopathy status at the time of ECMO initiation remained relevant throughout the ECMO course. Moreover, the timing of our Duplex ultrasound after decannulation renders it challenging to rule out pre-existing vascular complications. It is plausible that conditions like arterial dissection or arteriovenous fistula could have occurred at the time of cannulation but remained undetected until decannulation. The ultrasound window during ECMO treatment would likely have been insufficient to detect such complications due to the presence of the cannula. Finally, because Duplex ultrasound was not performed on all patients following decannulation, there is the potential for selection bias, possibly leading to an inflated estimation of complication rates. Nevertheless, Duplex ultrasound was conducted as routinely as feasible after decannulation. To address these limitations, future multicenter studies are needed to externally validate our findings, with a particular focus on the incidence of vascular complications and the factors influencing their occurrence across different institutions.
Arterial complications requiring either endovascular intervention or surgery frequently occur following successful weaning from VA-ECMO detected by Duplex ultrasound. DIC appears to be associated with an increased rate of arterial complications. Because acute limb ischemia after decannulation can contribute to in-hospital mortality, physicians should carefully monitor patients for the development of arterial complications.
None.
This study did not receive any specific funding.
This study was approved by the IRB of Samsung Medical Center (IRB no. 2020-10-102).
The datasets used and/or analyzed during the present study are available from the corresponding author upon reasonable request.
Please find supplementary file(s);
https://doi.org/10.1253/circj.CJ-24-0400
