Retrospective Analysis of 99 Patients With the Application of Extracorporeal Membrane Oxygenation in Fuwai Hospital

Abstract

The purpose of this study was to retrospectively summarize and analyze the data of 99 cases with the application of consecutive extracorporeal membrane oxygenation (ECMO) from December 2004 to August 2008 in Fuwai Hospital. The intersurface of the ECMO equipment system was completely heparin coated. All patients received venoarterial ECMO, and the activated clotting time was maintained between 120 and 180 seconds. The heparin dose was 5–20 U · kg⁻¹ · h⁻¹. Mean blood flow was 40–220 ml · kg⁻¹ · min⁻¹ during ECMO assisted period. The shortest ECMO time was 12 hours, and the longest was 504 hours (mean time, 119.45 ± 80.20 hours). Sixty patients (60.6%) weaned from ECMO successfully; 54 of them (84%) were discharged, and six died of postoperative complications. Thirty-nine patients could not wean from ECMO. Total discharge rate was 54.5%. ECMO is an effective mechanical assistant therapy for cardiac and pulmonary failure after cardiac surgery. Earlier use of ECMO for heart lung failure patients and avoidance of irreversible damage to the main organs are still the key points in the success of ECMO.

Extracorporeal membrane oxygenation (ECMO) is a technique to provide effective mechanical pulmonary and circulatory assistance for patients with various etiologies of cardiac failure refractory to conventional medical therapies. It has been successfully used as a bridge to myocardial recovery or cardiac transplantation. However, despite major advances in device technology and the intensive care of these severely ill patients, disproportionately high short-term mortality rates persist.^1,2 In this study, we retrospectively analyzed the charts of the 99 consecutive patients who received ECMO cardiopulmonary support between December 2004 and August 2008 in Fuwai Hospital.

Materials and Methods

Ninety-nine patients were treated with ECMO from December 2004 to August 2008, 46 infants and children and 53 adults including 66 men and 33 women. Age range of these patients was from 5 days to 76 years (mean age, 28.43 ± 26.25 years). Weight range was from 3.8 to 100 kg (mean weight, 42.28 ± 29.86 kg). Eighty-nine patients were postcardiac surgery; 10 were from cardiac internal medicine wards who received ECMO after cardiopulmonary resuscitation (CPR).

ECMO Indications

Venoarterial ECMO support was initiated under the following circumstances: 1) cardiopulmonary arrest (patient undergoing effective CPR on intubation); 2) failure to withdraw from cardiopulmonary bypass; 3) postoperative low cardiac output (CI < 2.0 · min⁻¹ · m⁻²); 4) severe hypoxemia (PaO₂/FiO₂ < 60%; and 5) miscellaneous conditions.

ECMO Equipments and Circuits

Medtronic Minimax Plus Oxygenator, Bio-Pump Plus Centrifugal Blood Pump and Bio-Probe Flow Transducer (Medtronics, Minneapolis, MN) and a heat exchanger were used in each patient. The whole circuit, including cannulae, tubing, oxygenator, and centrifugal pump, was treated with Carmeda BioActive Surface (tip to tip) (Carmeda, Upplands Vasby, Sweden). The probe for monitoring oxygen saturation and the venous negative pressure detector were mounted in the proper position in the ECMO system. An oxygen-air mixer (Sechrist Industries, Anaheim, CA) was used to ventilate the membrane oxygenator.

Cannulation

The venoarterial ECMO mode was used in all patients. In all pediatric patients, ECMO was initiated through the cannulae of the right atria and the ascending aorta through the thoracic incision. In one adult, ECMO was initiated through the cannulae of the right atria and the femoral artery, whereas the rest of the adults were initiated through the cannulae of the femoral vein and artery. When percutaneous femoral ECMO was instituted, an additional 16F intravenous needle casing was inserted distally into the femoral artery to prevent severe leg ischemia.

Patient Management During ECMO

During the ECMO period, nine patients were maintained in waking state without tracheal intubation, whereas other patients were sedated by morphine and midazolam to achieve patient comfort. Body temperature was maintained between 36°C and 37°C. Anticoagulation was provided by heparin through the ECMO circuit to keep the activated clotting times (ACT) between 120 and 180 seconds unless the patient had excessive postoperative bleeding; in those cases, lower ACT was maintained. Intravenous inotropic doses were decreased or stopped to alleviate myocardial oxygen consumption and promote myocardial recovery. When weaning from ECMO, vasoactive drugs were administered according to the patient’s hemodynamic state. According to the hemodynamics and blood gas results, mean ECMO flow was maintained at 40–220 ml · kg⁻¹ · min⁻¹), and an air-oxygen mixer was used for membrane oxygenator ventilation. To achieve mixed venous oxygen saturation more than 70%, PaO₂ was about 150 mm Hg. We maintained the fraction of oxygen of the air-oxygen mixer between 40% and 70%. Oxygen flow was gradually adjusted to meet the oxygen requirements of the patients. Except the nine patients without intubation, all patients received synchronized intermittent mandatory ventilation (SIMV). The ventilator was maintained at rest settings: peak airway pressure was <25 cm H₂O; tidal volume was 8–10 ml/kg; peak end-expiratory pressure was between 4 and 6 cm H₂O; ventilator rate was 10–30 breaths per minute; and FiO₂ was 0.30–0.60. Echocardiograph, hemodynamics, chest x-ray, and blood gas results were used to assess heart and lung function and ensure adequate systemic perfusion. At weaning, if the condition of patients improved to a satisfactory state, the flows were gradually reduced at hourly intervals during a 12–48-hour period, and the inotropic drugs were increased at the same time according to the hemodynamic monitor. The ACT was maintained at more than 200 seconds during the trials off ECMO to prevent clotting in the cannulas. Once the patients were hemodynamically stable on the minimal ECMO flow (10%–20% of total flow) with good recovery of myocardial contractility, evidenced by echocardiograph, patients were weaned from ECMO support.

Statistical Analysis

Statistical analysis was performed using SPSS 11.0 (SPSS Inc., Chicago, IL) software. The data were expressed as the arithmetic mean and the standard deviation. The Fisher’s exact test was used to compare several categorical variables. Statistical differences between two groups for continuous measurements were assessed using the independent-samples t-test. If the variance of the two collectively is unequal, rank-sum test was performed. Significance was accepted at the 5% level (p < 0.05).

Results

Extracorporeal membrane oxygenation was installed in 55 patients (55.5%) because of postoperative low cardiac output [20 underwent CPR during ECMO cannulation (10 of these 20 patients were postoperative; seven had end-stage cardiomyopathy, and three had coronary heart disease)]; in 15 patients (15.2%) because of low cardiac output and respiratory failure; and in nine patients (9.1%) because of respiratory failure. Figure 1 shows the indications of ECMO support of the different groups. The patients were divided into four groups by preoperative indications.

Figure 1.: Extracorporeal membrane oxygenation (ECMO) indications. 1, low cardiac output; 2, CPR; 3, low cardiac output and respiratory failure; and 4, respiratory failure.

ECMO Results

Extracorporeal membrane oxygenation support time was 12–504 hours, (mean time, 119.45 ± 80.20 hours). Sixty patients (60.6%, 60/99) were weaned off ECMO successfully, mean support time was 108.02 ± 59.25 hours; 54 patients (90.0%, 54/60) were discharged without severe complications; and six patients died after weaning off. Treatment was terminated in the remaining 39 patients because they could not be weaned off successfully. The total percentage of discharge from the hospital was 54.5%.

The patients were divided into five groups by age according to the relation to their indications for ECMO, as shown in Figure 2. In Figure 3, the patients were divided into three groups by weight. There were no significant differences between groups. Variables associated with outcomes of ECMO are shown in Table 1.

Figure 2.: Age distribution of patients with extracorporeal membrane oxygenation (ECMO). 1, infants; 2, preschool children; 3, children; 4, adult; and 5, senium. There was no significant difference between different groups.

Figure 3.: Weight distribution of patients with extracorporeal membrane oxygenation (ECMO). There was no significant difference between different groups.

Table 1: Variables Associated With Outcomes of ECMO (x̄ ± s )

ECMO Complications

The complications were grouped as follows: 1) postoperative bleeding and oozing [24 (24.2%)]; 2) renal dysfunction [11 (11.1%)]; 3) bloodstream infection [12 (12.1%)]; 4) thrombus [11 (11.1%)]; 5) central nervous system injury [6 (6.1%)]; 6) renal and hepatic dysfunction [5 (5.1%)]; 7) hemolysis [5 (5.1%)]; 8) gastrointestinal hemorrhage [2 (2.0%)]; 9) DIC [1 (1.0%)]; 10) heparin-induced thrombocytopenia [1 (1.0%)]; and 11) mechanical failure of ECMO circuit [16, (16.2%)]. Table 2 shows the ECMO complications and hospital survival.

Table 2: The ECMO Complications and Hospital Survival

Discussion

Extracorporeal membrane oxygenation has been used increasingly in recent years. About 3%–8% of the patients who undergo cardiac surgery need mechanical circulatory support.^3–5 It is important in the treatment of refractory postoperative low cardiac output, hypoxemia, arrhythmia, cardiac arrest, and failure of weaning off. Extracorporeal membrane oxygenation can also be an important adjunctive tool in the management of patients awaiting heart transplantation.⁶ In addition, ECMO may be a lifesaving measure by allowing an interval for return of native ventricular function in the majority of the patients with viral myocarditis.⁷

In our study, 89 patients (89.9%) received ECMO after cardiac surgery (65 for low cardiac output, 10 underwent CPR, nine for respiratory failure, and 15 for low cardiac output and respiratory failure). Ten nonsurgery patients received ECMO after CPR (seven patients because of end-stage cardiomyopathy and three because of coronary heart disease).

In all pediatric patients, ECMO was initiated through the cannulae of the right atria and ascending aorta through the sternotomy. In one adult, ECMO was initiated through the cannulae of the right atria and the femoral artery because of inadequate venous drainage, whereas in the rest of the adults, ECMO was initiated through the femoral vein and the artery cannulae. In our institution, the majority of our infant patients failed to wean from CPB, we usually switched to the ECMO circuit through the trans-thoracic cannulae, which were already in situ and support the patient with the chest open. In addition, VA access could provide sufficient flow. In older children and adults, access for extrathoracic VA ECMO was through the femoral vein and the artery cannulae. If venous drainage was not sufficient, our cannulae were placed in the right atria to achieve an adequate drainage.

To prevent ischemia of the lower extremity, we used a size of 16F intravenous needle casing for distal perfusion of the femoral artery, but there were still six patients who had lower limb ischemia. This indicated that even if extra cannulae were used to maintain distal limb perfusion, careful observation would be required to prevent limb ischemia.

During ECMO, most of our patients were sedated and perhaps anesthetized to facilitate ventilator management. Nine patients were maintained in waking state without trachea cannula. We prefer not to anesthetize patients, keeping them as awake and alert as possible.

In nonsurvival patients, the pre-ECMO mean arterial pressure (MAP) and the concentration of serum lactic acid were significantly higher than that of the discharged patients. The pre-ECMO MAP (60.5 ± 23.85 mm Hg) of the discharged group was much higher than that of nonsurvivors (41.7 ± 31.34 mm Hg) and the lactic acid concentration in pre-ECMO artery blood of the nonsurvivors (9.47 ± 2.35 mmol/L) was significantly higher than that of the discharged patients (6.87 ± 2.11 mmol/L). Cheung et al. reported similar results indicating that patients with evidence of significant tissue hypoperfusion, as manifested by a high-serum-lactate level, had a surprisingly higher mortality.⁸ So, early implementation of ECMO to improve organ perfusion before the development of end-organ damage could result in better outcomes. The mean ECMO support time in nonsurvivor group was 137.28 ± 10.58 hours and in the discharged group was 108.68 ± 58.98 hours. There were no significant differences between these two groups. From the study of 84 children with ECMO, Shah et al.⁹ demonstrated that only 5% survival rate was achieved in the patients with ECMO circulated more than 144 hours.

Bleeding is the most common complication of ECMO. According to the Registry Report 2004 of the Extracorporeal Life Support Organization (ELSO), the overall bleeding complication rate was 29%.¹⁰ In this study, we observed 22 cases (22.2%) of bleeding and oozing in 99 patients. However, there was no significance difference between the survivor and the nonsurvivor groups that was clearly due to our use of anticoagulation therapy during ECMO support. In our early cases with ECMO (before 2007), we observed a high incidence of bleeding complications. Therefore, we have changed our anticoagulation protocol. Systemic heparinization was not initiated until chest drainage was reduced to <50 mL/h. When the pump flow was high, lower ACT (120–160 seconds) was accepted. No thrombus was found. Our ACT time was shorter than most reports, which were 160–180 seconds.¹¹ One patient developed a thrombosis in the left atrium, despite successful laboratory anticoagulation. We assumed that this thrombosis was the consequence of large volume of venous drainage, which reduced the blood supply to the left atrium. So, the speed of blood flow decreased.

In addition, renal dysfunction, infection, thrombus formation, renal and hepatic failure, lower limb ischemia, hemolysis, and gastrointestinal hemorrhage are common complications in ECMO. Renal failure was the most significant risk factor in the nonsurvivors. This result was similar to Aparna et al.¹² who reported that multiple organ failure (MOF) was significantly related to mortality. MOF was not found in our analysis, which may be relevant to few organ failure cases.

Oxygenator plasma leakage is a common complication in ECMO management. ELSO reported 7.1%–16.7% of the positive rate.¹⁰ Sixteen oxygenators had to be exchanged because of serious leakage of plasma, which was associated with the oxygenator type, the transmembrane pressure, blood flow, and the degree of blood damage.^13–15 Plasma-free hemoglobin and its metabolite also attributed to the oxygenator failure.¹⁶ Because Medtronic Carmida Minimax Plus Oxygenator used hollow fibers with micropores on the membrane surface, there is a high possibility of plasma leakage. Motomura et al.^17,18 proposed a new type of silicone hollow fiber oxygenator that did not have plasma leakage after continuous use for more than 34 days.

The overall survival in our series was 54.5% (54/99 patients). Sixty patients (60.1%) weaned off ECMO successfully. By comparison, the survival rate was higher than was reported by ELSO¹⁰ in 2004. Hospital mortality was 45.5% (45 patients), of whom 39 patients died during ECMO support, and six additional patients died after ECMO was discontinued by family request. Other causes of death included incomplete surgical repair, severe postoperative complications such as renal failure and MOF. Lower-limb ischemia was significantly related to the cannulation site. Further research is required to determine the optimal methods to prevent, decrease, and treat ECMO complications.

Conclusion

On the basis of the our experience in the institution, ECMO therapy is now considered to be a valuable option for treatment of sudden cardiac arrest, postoperative low cardiac output, failure to wean from cardiopulmonary bypass, and as a bridge to heart transplantation among other indications. Earlier application of ECMO may decrease the complication rate and improve the overall survival rate.