Lung transplantation (LTX) has been established as a life-saving treatment for irreversible severe pulmonary disorders. The five-year survival rate of adult bilateral lung transplantation (BLTX) recipients is approximately 59% and that of adult single lung transplantation (SLTX) recipients is approximately 48% worldwide^[1]. In Japan, SLTXis indicated for 40%-65% of patients withend stage idiopathic pulmonary fibrosis or chronic obstructive pulmonary disease^[2], although the prognosis of SLTX is worse than that of BLTX^[1,3]. This is because the cadaveric donor is severely deficient; therefore, the indication of BLTX is very stringent in Japan^[2].

During LTX, extra-corporeal membrane oxygenation (ECMO) or cardiopulmonary bypass (CPB) is often used. These extracorporeal circulations (ECCs) are used for both hemodynamic and respiratory support in BLTX and mainly for respiratory support in SLTX. The criteria for induction of ECMO during LTX have been described previously^[4], which were made based on the clinical signs such as severe hypoxemia, right heart failure, and hemodynamic failure. On the other hand, the use of ECC is a risk factor for bleeding, infection^[5,6], and kidney dysfunction^[7]. Additionally,the use of ECC is thought to increase the mortality rate of LTX recipients^[4].

Therefore, avoidance of unnecessary ECC use can improve the prognosis of LTX recipients and also reduce healthcare cost^[8]. Moreover, if there is a factor predicting whether a LTX recipient needs ECC or not, this predictor can be used for pre-establishing an ECC in a patient actually needing ECC before a critical situation arises, which leads to safer anesthetic management. Therefore, identifying the predictive factor for the need for ECC in LTX recipients is important.

The present study was an exploratory study, focusing on respiratory factors among the factors to which ECMO is applied in SLTX recipients, in order to link it to a nation wide survey.

Materials and Methods

Study Design and Population

Ethical approval was obtained from the Ethics Committee of the Tohoku University Graduate School of Medicine (reference number: 2018-1-803). This retrospective observational study involve dadult patients aged >20 years who had undergone cadaveric SLTX at the Tohoku University Hospital from January 2010 to December 2018. Patients who had been transplanted in the past were excluded. Written informed consent was obtained from all patients enrolled in this study.

Anesthetic Management

In our center, an arterial line (Safedraw, Merit Medical Inc., South Jordan, UT) was placed at the right radial artery in all SLTX recipients before the induction of general anesthesia. General anesthesia was induced with propofol or midazolam with remifentanil, fentanyl, and rocuronium. After the intubation of a double lumen endobronchial tube, a transesophageal echocardiographic probe was inserted. General anesthesia was maintained with propofol, remifentanil, fentanyl, and rocuronium. A pulmonary artery catheter (Swan-Ganz catheter, Edwards Lifesciences, Irvine, CA) and a central venous catheter (SMAC plus, Nippon Covidien, Tokyo, Japan) were placed via the right internal jugular vein. Then, the arterial blood and mixed venous blood were analyzed (ABL 800 FLEX, Radiometer, Bronshoj, Denmark). Subsequently, anesthesiologists and surgeons evaluated respiratory and hemodynamic status and decided whether to apply ECMO (MERA Centrifugal Blood Pump System, Senko Medical Instrument, Tokyo, Japan) according to the following criteria by Ius F et al^[4]: 1) Hypercapnia (E_TCO₂> 60 or PaCO₂> 60), 2) Hypoxemia (SpO₂< 90% or PaO₂<60 Torr), 3) Circulatory failure (cardiac index < 2 L/min/m₂) and 4) right heart failure (pulmonary blood flow ratio > 0.9). After the induction of anesthesia is completed, left LTX surgery is performed in the right lateral position and right LTX is performed in the supine position. If the anesthesiologist recognizes above clinical criteria for hypercapnia, hypoxemia, circulatory failure, and right heart failure in a recipient who does not have ECMO, the operation is temporarily stopped and ECMO is urgently indicated.

Data Documentation

All study variables were retrieved from the electronic medical records system used at our center(Prime Gaia^®, Nihon Kohden, Tokyo, Japan). To ensure the confidentiality of the patients, no identifying information was recorded. We extracted clinical and demographic characteristics of each patient, as well as laboratory findings, administered drugs and fluids, and clinical events at the time of SLTX.

Endpoint

Induction of ECMO was regarded as the primary endpoint of the present study. Patients divided into two groups (ECMO use group and ECMO free group).

DataCollection

We collected the following data of the SLTX recipients as the factors associating with the ECMO use. Patient background: age; sex; primary disease; flow rate of supplemental oxygen; preoperative left ventricular ejection fraction and trans tricuspid pressure gradient; and ventilation-perfusion ratio of the dependent lung were collected from the electronic medical records. Perioperative data: arterial blood gas analysis during the period after the induction of general anesthesia to the start of surgery; vital sign data including mean arterial pressure, mean pulmonary artery pressure, central venous pressure, and mixed venous oxygen saturation at the insertion of the pulmonary artery catheter; and operative information including duration of surgery, blood loss, left or right side of implantation, ischemic interval of the graft, and types and duration of ECMO were collected from the electronic anesthesia records. And postoperative data; duration of nitric oxide inhalation and mechanical ventilation; and length of intensive care unit (ICU) and hospital stay were also collected from the electronic medical records.

Statistical Analysis

We analyzed the aforementioned factors in ECMO use and free group. Between-group comparisons were analyzed using the Fisher exact test or the Wilcoxon rank sum test. Factors that showed significant differences in univariate analysis were analyzed using multiple logistic regression analysis. Post hoc analysis was conducted to determine whether the existing dataset contained the appropriate number of cases to determine significant differences. Multicollinearity was assessed using the variance inflation factor^[9]. A variance inflation factor exceeding 10 is regarded as indicative of serious multicollinearity, and values greater than 4.0 may be a cause for concern. Conformity to a linear gradient was graphically verified, and polynomial or logarithmic transformations were performed when necessary. All analyses were performed using JMP^® 14.0 (SAS Institute Inc., Cary, NC). P values below 0.05 were considered to indicate statistical significance.

Results

All cadaveric single lung transplant recipients during the study period were eligible for the analysis. Eventually, we analyzed data from 43 patients who met our inclusion criteria. The characteristics of the recipients are shown in Table 1.

Table 1: Patient’sCharacteristics

Analyses were performed using Wilcoxon and Fisher’s exact tests. Results are presented along with standard deviations (±). *: P<0.05, **: P<0.01. V-A: veno-arterial, V-V: veno-venous, ECMO: extra-corporeal membrane oxygenation, ICU: intensive care unit; PaO₂:arterial oxygen tension, FIO₂: fraction of inspired oxygen, PaCO₂:arterial partial carbon-dioxide tension, HCO₃

^aArterial blood gas analysis data sampled during the period after the induction of general anesthesia to the start of surgery.

^bVital sign data at the insertion of a pulmonary artery catheter after induction of general anesthesia and before starting the operation.

^cOne case was switched from V-A ECMO to V-V ECMO during surgery.

Figure 1: Flow chart for the evaluation of eligible patients

Twenty-two patients received veno-arterial ECMO, 8 patients received veno-venous ECMO, and 1 patient received veno-arterial ECMO, which was switched to veno-venous ECMO. Total 33 (76.7%) patients were enrolled in the ECMO use group. While, 10 patients (23.3%) were enrolled in the ECMO free group. There were no patients who received ECMO before entering the operating room.

In the blood gas analysis before general anesthesia, ECMO free group was lower than ECMO use group in arterial partial carbon-dioxide tension (PaCO₂) (ECMO free group 37.9 ± 5.9 mmHg, ECMO use group 47.3 ± 5.9 mmHg, p = 0.014) and HCO₃^- (ECMO free group 24.2 ± 2.6 mmol/L, ECMO use group 28.8 ± 4.4 mmol/L, p = 0.007). On the other hand, after general anesthesia, PaCO₂ was not significantly different in two groups (ECMO free group 53.0 ± 11.3 mmHg, ECMO use group 69.4 ± 25.5 mmHg, p = 0.056), but ECMO free group was lower than ECMO use group in HCO₃^- (ECMO free group 26.2 ± 3.8 mmol/L, ECMO use group 30.2 ± 4.2 mmol/L, p = 0.011, Table 1). Meanwhile, there were not statistically significant differences between two groups in arterial oxygen tension (PaO₂) to fraction of inspired oxygen (F_IO₂) ratio (P/F ratio), ventilation-perfusion ratio of the dependent lung, and other preoperative blood gas analysis data.

In the vital sign data after induction of general anesthesia and before starting the operation, ECMO free group was higher than ECMO use group in mean arterial pressure (MAP) and diastolic arterial pressure (dBP) (MAP: ECMO free group 76.9 ± 10.0 mmHg, ECMO use group 69.0 ± 8.4 mmHg, p = 0.038. dBP: ECMO free group 61.1 ± 8.2 mmHg, ECMO use group 54.3 ± 7.0 mmHg, p = 0.038, Table 1).

A pneumothorax occurred in one patient of ECMO free group who developed to tension pneumothorax and required urgent thoracic drainage procedure during implantation. Eventually this patient did not require intra- and postoperative ECMO. And three patients required urgent induction of V-A ECMO intraoperatively, because of hemodynamic collapse.

We performed logistic regression analysis to assess the association between ECMO indication and the factors that were significantly different in univariate analysis (HCO₃^- and MAP, Figure 2). The AUROC was 0.866, and R₂ was 0.358. The cutoff value of HCO₃^- was 29.9 mmol/L (Sensitivity: 96.6%, Specificity: 75.0%). In post hoc analysis, the detection power was 0.248. There was no evidence of multicollinearity, since the variance inflation factor for independent variables in all models in Figure 2 was less than 4.0. The Odd ratio for ECMO indication was 1.47 for HCO₃^- (95%CI: 1.19-2.17, p = 0.015, Table 2).

Table 2: Risk Factors of ECMO Indication

Statistical analysis was logistic regression analysis. R²: 0.358, AUROC: 0.866. ECMO: extra-corporeal membrane oxygenation, MAP: mean arterial pressure, OR: odds ratio, CI: confidence interval, AUROC: area under of receiver operating characteristic curve.

Figure 2: Receiver operating characteristics curve for predicting ECMO indication

Results of the logistic regression analysis showed the following:R², 0.358 and AUROC, 0.866. Cut-off value of HCO₃^- was 29.9 mmol/L (sensitivity: 96.6%, specificity: 75.0%). ECMO: extra-corporeal membrane oxygenation

The distribution of HCO₃^- values in each group and the obtained cutoff values are shown in the Figure 3.

Figure 3: Distribution and the cutoff value of HCO₃^- before induction of general anesthesia

The cutoff value of HCO₃^- was 29.9 mmol/L. ECMO: extra-corporeal membrane oxygenation

In the outcomes, the ECMO use group had a longer operation time (549 ± 92 vs. 409 ± 40 min, p < 0.001), more blood loss (1488 ± 1642 vs. 265 ± 156 g, p = 0.001), and longer ICU days length of stay (12 ± 9 vs. 26 ± 40 days, p = 0.045, Table 1).

Discussion

In the present study, we found the potential predictors of ECMO use in SLTX recipients. The potential predictors were arterial PaCO₂ and HCO₃^- taken during the period after the induction of general anesthesia to the start of surgery. In other study factors, mean and diastolic arterial pressure also showed significantly difference which might indicate that arterial pressure was used as a reference. And we also found that large portion of SLTX recipients (79.5%) received ECMO, and the median duration of ECMO use was one day which suggests that the most SLTX recipients received ECMO were withdrawn the ECMO on the day of surgery. Additionally, the use of ECMO prolonged the surgery period, increased blood loss during surgery, and significantly extended the ICU days length of stay.

It is a big problem that the setting of mechanical ventilation greatly influences a PaCO₂. Meanwhile, HCO₃^- is regulated by metabolism and changes chronically. Therefore, the mechanical ventilation less likely affects the HCO₃^- immediately after the induction of general anesthesia, which could reflect the preoperative respiratory condition.

In this study, the cut-off value of HCO₃^- was determined to be 29.9 mmol/L in the ROC curve by logistic regression analysis using HCO₃^- and MAP as a covariate. Although this method should be originally examined in a prospective comparative study, it was performed because it was considered clinically significant to determine a provisional cutoff value, since there are very few cadaveric lung transplantations in Japan.

Henderson–Hasselbalch equation in acid-base equilibrium of blood is known as follows^[10].

pH = 6.1 + log ([HCO₃^-]/(0.03 × PaCO₂)).

In case a patient is in a stable condition and the pH of arterial blood is 7.4, the equation transforms into as follows.

PaCO₂ = [HCO₃^-]/0.60.

Assigning the cutoff value of this study, that is, 29.9 mmol/L to HCO₃^- of this equation, the value of PaCO₂ is calculated to 49.8 Torr. This PaCO₂ value also might mean the preoperative cutoff value of inducing ECMO. The difference between HCO₃^- in ECMO group and non-ECMO group is 4 mmol / L, which seems to be small, but when assigned to PaCO₂, it becomes a difference of 24 Torr, which is considered to be a large difference.

In our case, two patients required emergency induction of intraoperative V-A ECMO, and the arterial HCO₃^- collected before general anesthesia was 36.8 and 41.1 mmol / L, respectively. These were higher than the cutoff value of 29.9 mmol / L. In these cases, their PaCO₂ had elevated beyond 120 mmHg within a few hours of the start of mechanical ventilation and their systolic pulmonary artery pressure had also elevated beyond 60 mmHg. After that, their systemic arterial pressure had dropped below 60 mmHg and their responsiveness to vasopressor had been decreased. They had developed to acute right heart failure and subsequent hemodynamic collapse. In this situation, we had to induce V-A ECMO very urgently. In Figure 2, in the ECMO free group, all patients had HCO₃^- below the cut-off value, so it may be recommended not to use ECMO in SLTX patients with preoperative HCO₃^- below 29.9.

This study has several limitations that can be addressed in future studies to improve clinical relevance. Because this was a single-center retrospective analysis with the assumption that a large multicenter study would be conducted in the future, we were unable to rule out potential sources of bias, such as the selection of study variables, the primary disease being studied, and factors related to the surgical technique. Furthermore, the power in the multivariate analysis was not sufficient at 0.248. A multicenter study of approximately 270 patients would be sufficient to obtain sufficient reliability. In addition, at the time of the study period, the criteria for ECMO inclusion were not sufficiently rigorous. Future studies should also investigate the influence of hemodynamic variables on ECMO induction in BLTX patients.

Conclusion

Evaluation of arterial carbon dioxide tension and bicarbonate concentration could predict the demand of ECMO use in SLTX. Additionally, the preoperative evaluation of arterial carbon dioxide tension and bicarbonate concentration has possibility to inhibit unnecessary ECMO use and increase the safety of anesthetic management by previous preparation of ECMO for candidate.

Acknowledgement

This work was supported by JSPS KAKENHI Grant Number 19K18343

References

1. 1. Lund, L.H., Khush, K.K., Cherikh, W.S., et al. The Registry of the International Society for Heart and Lung Transplantation: Thirty-fourth Adult Heart Transplantation Report-2017; Focus Theme: Allograft ischemic time. (2017) J Heart Lung Transplant 36(10):1037-1046.

   PubMed│CrossRef│Others

2. 2. Network JOT. Lung transplant applicant Tokyo, Japan: Japan Organ Transplant Network; 2019 [updated February 28, 2019.

   PubMed│CrossRef│Others

3. 3. Transplantation TJSf. Fact book 2018 of organ transplantation in Japan Tokyo, Japan: The Japan Society for Transplantation; 2018.

   PubMed│CrossRef│Others

4. 4. Ius, F., Sommer, W., Tudorache, I., et al. Five-year experience with intraoperative extracorporeal membrane oxygenation in lung transplantation: Indications and midterm results. (2016) J Heart Lung Transplant 35(1):49-58.

   PubMed│CrossRef│Others

5. 5. Wong, J.K., Melvin, A.L., Joshi, D.J., et al. Cannulation-Related Complications on Veno-Arterial Extracorporeal Membrane Oxygenation: Prevalence and Effect on Mortality. (2017) Artif Organs 41(9):827-834.

   PubMed│CrossRef│Others

6. 6. Bisdas, T., Beutel, G., Warnecke, G., et al. Vascular complications in patients undergoing femoral cannulation for extracorporeal membrane oxygenation support. (2011) Ann Thorac Surg 92(2):626-631.

   PubMed│CrossRef│Others

7. 7. Kumar, A.B., Suneja, M. Cardiopulmonary bypass-associated acute kidney injury. (2011) Anesthesiology 114(4):964-970.

   PubMed│CrossRef│Others

8. 8. Hayanga, J.W.A., Shigemura, N., Aboagye, J.K., et al. ECMO Support in Lung Transplantation: A Contemporary Analysis of Hospital Charges in the United States. (2017) Ann Thorac Surg 104(3):1033-1099.

   PubMed│CrossRef│Others

9. 9. Robert, M.O. A Caution Regarding Rules of Thumb for Variance Inflation Factors. (2007) Quality & Quantity 41(5):673-690.

   PubMed│CrossRef│Others

10. 10. Hasselbalch, K.A. The calculation of the hydrogen number of the blood from the free and bound carbonic acid of the same, and the oxygen binding of the blood as a function of the hydrogen number. (1916) Biochem Z. 78:112- 144.

    PubMed│CrossRef│Others