After oesophagectomy, postoperative complications occur in 30 to 50% of patients.1,2 Neoadjuvant therapy with either chemotherapy or chemoradiotherapy improves long-term survival after oesophagectomy,3 but the addition of radiotherapy may impose adverse effects on the perioperative course compared with when chemotherapy is given alone. However, the two previous randomised trials that have compared chemotherapy with chemoradiotherapy did not include details on the postoperative course.4,5 The recently completed Swedish–Norwegian NeoRes trial compared neoadjuvant chemotherapy and chemoradiotherapy and reported a higher degree of histological response after chemoradiotherapy, but no difference in 3-year survival.6 As regards long-term survival, the oncological benefit from the addition of radiotherapy may have been countered by the significant increase in mortality because of serious adverse events and postoperative complications in the chemoradiotherapy arm during the first year after randomisation. In addition, the postoperative complications were more severe after chemoradiotherapy.7 These data suggest that the addition of radiotherapy may have a significant adverse effect on the lungs and heart, which are in the radiation field.
We have previously observed negative effects on preoperative myocardial function as well as an increase in N-terminal pro-B-type natriuretic peptide (NT-proBNP) in association with neoadjuvant chemoradiotherapy.8 These observations are in accordance with those obtained by others.9–11 Pre and postoperative levels of NT-proBNP have been suggested as prognostic indicators of postoperative cardiac morbidity.12–15
The aim of this study was to compare the effects of neoadjuvant chemotherapy and chemoradiotherapy on perioperative cardiac function in patients undergoing thoracoabdominal oesophagectomy. The patients were in a cohort from the NeoRes trial.6 We hypothesised that, in the perioperative setting, neoadjuvant chemoradiotherapy would impair cardiac function more than neoadjuvant chemotherapy. The primary outcome in this study was the effect of the neoadjuvant treatment on stroke volume index (SVI) during the perioperative period. SVI was chosen as, given a similar preload (by fluid administration), it is the variable best describing cardiac contractility that was clinically feasible to measure in this setting. Secondary outcomes were the effects of treatment on the oxygen delivery index (DO2I), echocardiography and biochemical markers. DO2I was chosen as it represents the overall end result of cardiovascular function. Echocardiography and biochemical markers were included to detail cardiac function using additional methods.
Patients and methods
Patients recruited for this study represent a cohort from those enrolled in the NeoRes trial, a multi-centre randomised, nonblinded trial of neoadjuvant chemotherapy vs. chemoradiotherapy for cancer in the oesophagus or gastrooesophageal junction (NeoRes 2006–2013; EudraCTnr 2006–001785-16). Patients included were those scheduled for surgery at the Karolinska University Hospital who, from 2009 onwards, were subjected to an extended protocol involving haemodynamic studies. A detailed description of the inclusion and randomisation processes for the NeoRes trial has been published elsewhere.6 An additional inclusion criterion for the extended protocol was planned thoracoabdominal surgery at the Karolinska University Hospital, Huddinge.
The NeoRes study protocol was approved by the regional ethics committee (EPN Stockholm 2006/738-32 and 2008/1822-32) and registered at the US National Institute of Health (ClinicalTrials.gov NCT01362127). For the extended protocol, additional ethical approval was obtained (Amendment 2008/1822-32) on 13 January 2009 (Professor H. Glaumann). All patients received written and oral information, and were then included in the current study after signed informed consent had been obtained.
The neoadjuvant treatment regimens have been detailed elsewhere.6 Briefly, patients received three 21-day cycles of chemotherapy: Cisplatin 100 mg m−2 was given on day 1 and 5-fluorouracil 750 mg m−2 day−1 was given on days 1 to 5. Thereafter, 16 days were allowed for recovery before the next cycle commenced. In the chemoradiotherapy group, radiotherapy was added during cycles two and three using a computed tomography-based three-dimensional planning system: 40 Gy, with a multiple field technique (2 Gy once daily in 20 fractions), 5 days per week for 4 weeks, starting week 1 of second cycle. During radiotherapy, the planning target volume receiving more than 95% of planned radiation dose as well as the planning target volume receiving 95 to 105% of the planned dose was measured. The percentage of the heart volume receiving 10 Gy or more as well as 30 Gy or more was calculated.
Surgery and anaesthesia
Perioperative treatment was the same in both groups. Surgery was performed via a thoracoabdominal approach according to either Ivor Lewis (right thoracotomy and laparotomy) or McKeown (right thoracotomy followed by laparotomy and neck incision with cervical anastomosis), depending on the tumour location. All patients received bilateral active thoracic drainage and a jejunal catheter for postoperative enteral nutrition. Preoperatively, an epidural catheter was inserted between what was estimated to be the sixth–eighth intervertebral space and 1 ml of fentanyl (50 μg ml−1) was given as a bolus through the epidural catheter. An epidural infusion [bupivacaine (1 mg ml−1), epinephrine (2 μg ml−1) and fentanyl (2 μg ml−1)] was then commenced at a rate of 10 to 15 ml h−1. Induction of anaesthesia was with propofol 2 to 3 mg kg−1 and fentanyl 2 μg kg−1, and atracurium 0.6 mg kg−1 was used for muscle relaxation before intubation. Further doses of atracurium (0.1 to 0.2 mg kg−1) were given as required. Sevoflurane was administered for maintenance of anaesthesia. All patients had their tracheas intubated with a double lumen bronchial tube, the position of which was verified by bronchoscopy. Single-lung ventilation was used during the thoracic part of the operation. A norepinephrine infusion (40 μg ml−1) was used to maintain mean arterial pressure at 60 to 70 mmHg. Fluids were given as follows: glucose (25 mg ml−1) with electrolytes (Braun, Melsungen, Germany) at 1 ml kg−1 h−1, Ringer's acetate (Baxter, Viaflo, Deerfield, Illinois, USA) at 2 ml kg−1 h−1 and poly-hydroxyethyl-starch 130/0.4 (Voluven 60 mg ml−1, Fresenius Kabi, Homburg, Germany) at 2 ml kg−1 h−1. Hypovolemia was treated with poly-hydroxyethyl-starch and blood products as deemed appropriate by the anaesthetist. All study participants had the double lumen endotracheal tube removed at the end of surgery and were transferred to the postoperative recovery ward. All except three patients in the chemotherapy group had an open thoracoabdominal oesophagectomy with a gastric tube reconstruction. One patient had a concomitant total gastrectomy with a distal pancreatosplenectomy and a Roux-en-Y oesophagojejunostomy. One patient had the abdominal part of the operation done via laparoscopy and one patient had a concomitant pulmonary resection because of emphysema.
LiDCO and LiDCOplus
Hemodynamic measurements were performed using the LiDCOplus (LiDCO Ltd, London, United Kingdom).16,17 Calibration was performed according to the manufacturer's instructions, with the exception that three instead of two calibration points were used for optimal accuracy.18 A central venous line was used for injection of lithium indicator for the dilution studies.19
With the patient resting and without the use of sedative drugs, haemodynamic measurements after fluid optimisation were recorded in the preoperative ward before surgery. During surgery, haemodynamic measurements without fluid optimisation were recorded at the following time points: after 2 h of abdominal surgery, 1 h after the start of one-lung ventilation, and after closure of the thoracic wound. On the third postoperative day (POD 3), haemodynamic measurements after fluid optimisation were repeated. Fluid optimisation was performed using boluses of 3 ml kg−1 (2.5 ml kg−1 if BMI > 30) poly-hydroxyethyl-starch 130/0.4, which was repeated if stroke volume (SV) increased more than 10%. Values were recorded when they had been stable for more than 1 min. In the case of a borderline SV increase, a passive leg raise was performed and, if positive (SV increased >10%), additional fluids were given.20 During surgery, SV values were recorded when they had been stable for more than 1 min. The monitor was covered and the data were not available to the attending anaesthetist. To further standardise the measurements preoperatively, patients who received laxatives the night before surgery (in case there was a need to use the colon as a conduit) also received a supplement of with 1000 ml Ringer's acetate intravenously overnight. Postoperatively, pain was assessed by asking the patients to mark a 10 cm long visual analogue scale, and pain was treated as required until the visual analogue scale was below 4. Supplemental oxygen was given to maintain arterial oxygen tension more than 10 kPa. If haemoglobin (Hb) levels were less than 80 g l−1, packed red blood cells were given as boluses until Hb reached a level of more than 80 g l−1 before haemodynamic measurements were performed.
Echocardiography was performed before surgery some 4 to 6 weeks after neoadjuvant therapy, and repeated after fluid optimisation on POD 3 The analyser was blinded to study group allocation. Details of the method used have been published elsewhere.8 Owing to problems with poor image quality because of mediastinal dissection and wound dressings, the examinations from POD 3 could not be analysed.
N-terminal pro-B-type natriuretic peptide, troponin T
Venous blood samples were collected in EDTA tubes on admission for surgery, directly after surgery, and on the mornings of PODs 1 to 3. An electrochemiluminescence immunoassay with Modular Analytics E170 (Roche Diagnostics, Mannheim, Germany) was used for the analyses. All analyses were performed in the accredited laboratory for clinical chemistry (ISO 15189) within the Karolinska University Hospital.
In-hospital morbidity and mortality was recorded as follows: the composite outcomes major adverse cardiac events and postoperative pulmonary complications were recorded daily by the research team according to European Society of Anaesthesiology-European Society of Critical Care Medicine) guidelines.21 Surgical complications were defined as any complication directly linked to the surgical procedure (e.g. anastomotic leakage, chylothorax or similar types of complications). Sepsis was defined according to the 2001 International Sepsis Definitions Conference.22
As patients included were selected from a larger multi-centre study, no power calculation was feasible. Given the high rate of failure to complete neoadjuvant treatment and surgery (because of disease progression and death), before the study began, we had decided to analyse patients completing resection as per protocol, as our aim was to analyse perioperative haemodynamic function. A linear mixed model for repeated measures was used for analysis. The model was used to analyse the trend effect (the change between the measurements over time), within-group change (the difference between measurements over time in the separate groups) and the interaction effect (the difference in change between groups). This was defined as the basic model. In addition, the influence of different covariates was analysed. An autoregressive covariance structure was used for data with multiple measurements, and an unstructured covariance structure was used for data with two measurements. Data with multiple measurements were tested for differences between the groups at the first measurement (preoperatively), as the groups were not considered to be equal based on the small sample size (a cohort of a larger randomised clinical trial) and on them having received different neoadjuvant treatments. Several mixed models analyses were performed, and in each analysis, one of a set of covariates was added to the basic model to test for any influence on the interaction effect. The set of covariates included sex, age, BMI, Physiological and Operative Severity Score for the Enumeration of Mortality and Morbidity, American Society of Anesthesiologists classification, hypertension, ischaemic heart disease, smoking, chronic obstructive pulmonary disease, anaemia, diabetes and alcohol abuse. Study participant characteristics and postoperative complications were analysed using a two-tailed Mann–Whitney U test and a two-tailed Fisher's exact test as appropriate. The Holm–Šidàk correction was used in post hoc testing. Linear mixed models were analysed with SAS 9.4 (SAS Institute Inc., Cary, North Carolina, USA). For all other statistics, Statistica 10 (StatSoft Inc., Dell, Texas, USA) was used. Data are presented as median (range) unless otherwise stated.
The NeoRes trial was closed in March 2013 when the planned 181 patients had been recruited. After starting this sub-study in January 2009, 55 patients were randomised into the NeoRes trial in Stockholm; of these, 41 were scheduled for thoracoabdominal surgery at Karolinska University Hospital at Huddinge (Fig. 1). Thirty-one patients completed surgery as per protocol and were analysed (chemotherapy n = 17, chemoradiotherapy n = 14). The chemoradiotherapy group was found to be older, 66 (56 to 75) vs. 60 (51 to 71) years (P = 0.03 and to have higher BMI 26 (21–34) vs. 23 (18 to 33) kg m−2 (P = 0.04), otherwise, the groups were well matched (Table 1). Age, sex and tumour grade were similar to the whole NeoRes study group.6 Six study participants, (two in the chemoradiotherapy group) did not complete neoadjuvant treatment because of adverse events. One study participant had a reduced radiation dose because of thoracic pain.
The intraoperative aspects were similar between the groups. Operating time 438 (302 to 640) vs. 440 (363 to 603) min (P = 0.93). Median blood loss was 900 (100 to 1900) vs. 690 (350 to 1300) ml (P = 0.86) and the number of patients needing transfusion was seven vs. eight (P = 0.51) in the chemoradiotherapy and chemotherapy groups, respectively. The volumes of fluids given during the operation were also similar. Crystalloid volume 36.8 (13 to 43) vs. 30.2 (12 to 60) ml kg−1 (P = 0.34) and colloid volume 29.1 (14 to 49) vs. 30.1 (16 to 46) ml kg−1 (P = 0.61) for the chemoradiotherapy and chemotherapy groups, respectively. There were no significant differences in the overall number of complications between the groups. Pulmonary morbidity dominated the complications, affecting 93 and 65% of the patients in the chemoradiotherapy and chemotherapy groups, respectively (P = 0.09, Table 2). Respiratory insufficiency (arterial oxygen concentration; fraction of inspired oxygen ratio <40 kPa) was the pre-dominant postoperative complication, affecting 79 vs. 65% of the patients (P = 0.46). Atrial fibrillation was the most common cardiovascular complication, with an incidence of 29 vs. 24% in the respective study groups (P = 1.0). One patient (6%) in the chemotherapy group had a myocardial infarction (MI). Three patients in each group developed sepsis. One patient in the chemoradiotherapy group died in hospital from multi-organ failure following an anastomotic leak.
SVI was similar between the groups during the study period (P = 0.27). In the preoperative ward before surgery, study participants in the chemoradiotherapy group had significantly lower cardiac index (CI); 2.9 (2.6 to 3.2) vs. 3.4 (3.1 to 3.8) l min−1 m−2 (P = 0.03), compared with the chemotherapy group. Mean CI remained lower in the chemoradiotherapy group during surgery but the interaction effect did not reach significance (P = 0.06). All haemodynamic variables changed significantly during the study period (within-group change), but there was no significant interaction effect between the groups. Adding the co-variable age to the model did not affect the P value for the interaction effect of SVI. Age did affect the interaction effect for other variables studied (CI, P = 0.06 to P = 0.10) but it did not change the interpretation of the interaction effect. Nor did any of the other co-variables tested have a significant impact on the results. Haemodynamic data are presented in Table 3 and Fig. 2a to f. During surgery, CI and DO2I decreased, reaching a minimum during the thoracic part of the procedure. There was no major change in Hb levels or SVR during the operation. On POD 3 SVI, heart rate, CI and DO2I had increased above baseline in both groups, whereas SVR and Hb decreased.
One patient in each study group declined to participate on POD 3. In five patients (three in the chemoradiotherapy group), measurements were not possible on POD 3 because of difficulties in obtaining adequate arterial lines. Epidural pain relief with the same mixture as used during surgery was ongoing in all patients on POD 3: dose rates ranged from 6 to 15 ml h−1. Five patients (three in the chemoradiotherapy group) were also receiving norepinephrine, with dose rates ranging from 0.01 to 0.04 μg kg−1min−1. Fluid administration was similar in both groups during the study period. The highest fluid volume was administered on the day of surgery, though mean fluid volume remained more than 4000 ml day−1 in both groups until POD 3 (data not shown).
N-terminal pro-B-type natriuretic peptide and troponin T
Plasma NT-proBNP levels were similar between the groups before surgery. The levels increased immediately after surgery and continued to increase throughout the study period. Both groups displayed a similar pattern with detectable differences between them (P = 0.94) (Table 3, Fig. 3a).
Troponin T levels were low in both groups throughout the study period and did not change over time (Fig. 3b). One patient in the chemotherapy group had a MI and was excluded from troponin T analysis, as representing an extreme outlier. None of the remaining patients were clinically suspected of having a MI.
In this study, we were unable to demonstrate that the addition of radiotherapy to neoadjuvant chemotherapy was followed by a significant impairment of cardiac function during the perioperative period. Neither SVI, nor any of the other haemodynamic variables analysed, changed differently between the groups over time. Also, and more importantly, we found no signs of impaired circulatory adaptation to surgical stress. We did find a decreased CI preoperatively in the chemoradiotherapy group (3.4 vs. 2.9 l min−1 m−2, P = 0.03), which is in accordance with our previously published echocardiography data.8 But, otherwise, the preoperatively assessed mean haemodynamic values were within normal limits in both groups, with no significant differences. Mean Hb values were low in both groups preoperatively (chemoradiotherapy 107 g l−1, chemotherapy 105 g l−1), probably as a result of chemotherapy-induced bone marrow depression. Both groups behaved similarly during surgery with a decrease in all haemodynamic parameters, with the lowest values generally attained during the thoracic part of the operation, which corresponds well to what has been reported previously in patients without neoadjuvant treatment.23 On POD 3, compared with baseline, both groups displayed a similar hyperdynamic circulation with a decreased SVR coupled with increased SVI, CI, DO2I and heart rate. This indicates that, when exposed to the postoperative stress, both groups are able to increase their cardiac performance to the same extent irrespective of which type of neoadjuvant therapy has been given.
NT-proBNP levels increased steadily over time, reaching levels normally suggestive of heart failure, and which might pre-dispose patients to postoperative cardiac events and mortality, as has been suggested by Rodseth et al.15 These events may originate from the postoperative inflammatory response and/or as a result of fluid overload. In a post hoc analysis, we found that NT-proBNP levels on POD 3 were higher in patients who required care in the ICU (P = 0.03) but not in those who developed cardiovascular complications (P = 0.06). Importantly, the profiles of NT-proBNP and troponin T were similar between the groups during the study period.
Postoperative pulmonary complications were slightly overrepresented (P = 0.09) in patients given chemoradiotherapy but fewer previous or ongoing smokers were included in the chemoradiotherapy group (14% as compared with 47% in chemotherapy group, P = 0.07) This trend may be something to take note of as, had there been similar smoking histories in the two groups, this difference in postoperative pulmonary complications between the groups may have been a statistically significant difference. These data would suggest that until a clear survival benefit of chemoradiotherapy compared with chemotherapy for cancer of the oesophagus or gastrooesophageal junction has been documented, chemoradiotherapy should be carefully evaluated in patients with significant pulmonary comorbidities.
The study's main limitation is the small sample size, representing a cohort of patients consecutively enrolled from a larger, multi-centre, randomised study. This introduces a risk for random error, which then reduces the statistical power. A quarter of the randomised patients were not included in the analysis because of the fact that they did not complete surgery. Accordingly, we chose to analyse as per protocol rather than intention to treat. This was justified as our aim was to describe the perioperative haemodynamics in the two groups, thus allowing for relevant hypothesis generation for future clinical trials. We also had problems with missing data: around one-fifth of the haemodynamic measurements on POD 3 were missing. Such missing data represent a more general problem: burdening normal levels of clinical staff with the additional data gathering required for this type of complex clinical research protocol, in patients already undergoing surgical procedures demanding a high level of care. However, to handle the missing data, we adopted a linear mixed model for the statistical analyses. It was not feasible to blind study group allocation during the study. However, study group allocation was blinded during echocardiography analyses. The imbalance in the age of the patient groups could affect our results. Cardiac function is known to decrease with age, but evidence is lacking to show a significant and clinically relevant decrease between 60 and 66 years of age.24,25 In addition, we did not observe any difference between the groups in exercise capacity before neoadjuvant treatment (data not shown). Adjusting the model for age rendered no change in the results.
The current results, although limited by sample size and missing data, suggest that compared to chemotherapy alone neoadjuvant chemoradiotherapy, as given in the present study for cancer of the oesophagus or gastrooesophageal junction, does not impose a clinically relevant negative cardiovascular effect. The marginal negative effect on baseline cardiac function elicited by the addition of radiotherapy had no clinically significant effect when patients’ haemodynamics were challenged by surgery.
Acknowledgements relating to this article
Assistance with the study: the authors thank RN Berit Sunde for administrative assistance, laboratory technician Maria Westerlind for performing the echocardiography exams, and lecturer Magnus Backheden from the Unit for Medical Statistics, Karolinska Institutet for discussions and assistance with statistical analyses.
Financial support and sponsorship: this study was supported by grants from the Swedish Medical Society and Stockholm County Council.
Conflicts of interest: none.
Presentation: preliminary data were presented as an oral presentation at the Swedish Association for Anaesthesia and Intensive Care annual meeting September 2015.
1. van Hagen P, Hulshof MC, van Lanschot JJ, et al. Preoperative chemoradiotherapy for esophageal or junctional cancer. N Engl J Med
2. Mariette C, Piessen G, Triboulet JP. Therapeutic strategies in oesophageal carcinoma: role of surgery and other modalities. Lancet Oncol
3. Sjoquist KM, Burmeister BH, Smithers BM, et al. Survival after neoadjuvant chemotherapy or chemoradiotherapy for resectable oesophageal carcinoma: an updated meta-analysis. Lancet Oncol
4. Stahl M, Walz MK, Stuschke M, et al. Phase III comparison of preoperative chemotherapy compared with chemoradiotherapy in patients with locally advanced adenocarcinoma of the esophagogastric junction. J Clin Oncol
5. Burmeister BH, Thomas JM, Burmeister EA, et al. Is concurrent radiation therapy required in patients receiving preoperative chemotherapy for adenocarcinoma of the oesophagus? A randomised phase II trial. Eur J Cancer
6. Klevebro F, Alexandersson von Dobeln G, Wang N, et al. A randomized clinical trial of neoadjuvant chemotherapy versus neoadjuvant chemoradiotherapy for cancer of the oesophagus or gastro-oesophageal junction. Ann Oncol
7. Klevebro F, Johnsen G, Johnson E, et al. Morbidity and mortality after surgery for cancer of the oesophagus and gastro-oesophageal junction: a randomized clinical trial of neoadjuvant chemotherapy vs. neoadjuvant chemoradiation. Eur J Surg Oncol
8. Lund M, von Dobeln G, Nilsson M, et al. Effects on heart function of neoadjuvant chemotherapy and chemoradiotherapy in patients with cancer in the esophagus or gastroesophageal junction inverted question mark a prospective cohort pilot study within a randomized clinical trial. Radiat Oncol
9. Erven K, Jurcut R, Weltens C, et al. Acute radiation effects on cardiac function detected by strain rate imaging in breast cancer patients. Int J Radiat Oncol Biol Phys
10. Hatakenaka M, Yonezawa M, Nonoshita T, et al. Acute cardiac impairment associated with concurrent chemoradiotherapy for esophageal cancer: magnetic resonance evaluation. Int J Radiat Oncol Biol Phys
11. Tripp P, Malhotra HK, Javle M, et al. Cardiac function after chemoradiation for esophageal cancer: comparison of heart dose-volume histogram parameters to multiple gated acquisition scan changes. Dis Esophagus
12. Cuthbertson BH, Amiri AR, Croal BL, et al. Utility of B-type natriuretic peptide in predicting perioperative cardiac events in patients undergoing major noncardiac surgery. Br J Anaesth
13. Karthikeyan G, Moncur RA, Levine O, et al. Is a preoperative brain natriuretic peptide or N-terminal pro-B-type natriuretic peptide measurement an independent predictor of adverse cardiovascular outcomes within 30 days of noncardiac surgery? A systematic review and meta-analysis of observational studies. J Am Coll Cardiol
14. Mahla E, Baumann A, Rehak P, et al. N-terminal pro-brain natriuretic peptide identifies patients at high risk for adverse cardiac outcome after vascular surgery. Anesthesiology
15. Rodseth RN, Biccard BM, Chu R, et al. Postoperative B-type natriuretic peptide for prediction of major cardiac events in patients undergoing noncardiac surgery: systematic review and individual patient meta-analysis. Anesthesiology
16. Jonas MM, Tanser SJ. Lithium dilution measurement of cardiac output and arterial pulse waveform analysis: an indicator dilution calibrated beat-by-beat system for continuous estimation of cardiac output. Curr Opin Crit Care
17. Kurita T, Morita K, Kato S, et al. Comparison of the accuracy of the lithium dilution technique with the thermodilution technique for measurement of cardiac output. Br J Anaesth
18. Cecconi M, Dawson D, Grounds RM, Rhodes A. Lithium dilution cardiac output measurement in the critically ill patient: determination of precision of the technique. Intensive Care Med
19. Jonas MM, Kelly FE, Linton RA, et al. A comparison of lithium dilution cardiac output measurements made using central and antecubital venous injection of lithium chloride. J Clin Monit Comput
20. Jabot J, Teboul JL, Richard C, Monnet X. Passive leg raising for predicting fluid responsiveness: importance of the postural change. Intensive Care Med
21. Jammer I, Wickboldt N, Sander M, et al. Standards for definitions and use of outcome measures for clinical effectiveness research in perioperative medicine: European Perioperative Clinical Outcome (EPCO) definitions: a statement from the ESA-ESICM joint taskforce on perioperative outcome measures. Eur J Anaesthesiol
22. Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Crit Care Med
23. Zou YB, Yan H, Liu XH, et al. Lateral position could provide more excellent hemodynamic parameters during video-assisted thoracoscopic esophagectomy for cancer. Surg Endosc
24. Kawel-Boehm N, Maceira A, Valsangiacomo-Buechel ER, et al. Normal values for cardiovascular magnetic resonance in adults and children. J Cardiovasc Magn Reson
25. EchoNo RC, EchoNo RENRM-AotLHC. Ethnic-specific normative reference values for echocardiographic LA and LV size, LV mass, and systolic function: the EchoNoRMAL Study. JACC Cardiovasc Imaging