This randomized clinical trial assesses the effect of total intravenous anesthesia vs inhalational anesthesia on postoperative pulmonary complications in patients undergoing microvascular reconstruction for head and neck cancer.
Key Points
Question
Do different anesthesia methods affect the postoperative pulmonary complication rate in patients undergoing free flap surgery?
Findings
In this randomized clinical trial that included 78 adults, the postoperative pulmonary complication rate was 14.3% in the total intravenous anesthesia (TIVA) group vs 40.0% in the inhalation group. The TIVA group had a relatively stable pattern of hourly hemodynamic parameters throughout the surgery.
Meaning
In free flap surgery of head and neck cancer, TIVA resulted in less hemodynamic fluctuation compared with inhalational anesthesia.
Abstract
Importance
Free flap surgery is a lengthy procedure with massive tissue destruction and reconstruction, which makes postoperative pulmonary complications (PPCs) a noticeable issue among patients with head and neck cancer. Propofol-based total intravenous anesthesia (TIVA) has better survival outcomes than inhalational anesthesia (INH) in several types of cancer surgery. A previous retrospective study found that patients in the TIVA group had a lower PPC rate, which may be correlated with a lower intraoperative fluid requirement. We hypothesize that the protective effect remains among patients undergoing free flap surgery for head and neck cancer in a prospective and goal-directed fluid therapy setting.
Objective
To assess the effect of TIVA vs INH on PPCs in patients undergoing microvascular reconstruction for head and neck cancer.
Design, Setting, and Participants
This prospective, 2-arm, randomized clinical trial was conducted at a tertiary hospital in Taiwan; a total of 78 patients 18 years and older with American Society of Anesthesiologists physical status classification 1 to 3 who were scheduled for elective free flap surgery under general anesthesia were included. The trial started in October 2017, completed in October 2019, and finished analysis in January 2022.
Interventions
Patients were enrolled and randomized to the TIVA or INH group. All patients received goal-directed fluid therapy and hemodynamic management if they had a mean arterial pressure (MAP) below 75 mm Hg or a reduction of 10% from baseline MAP.
Main Outcomes and Measures
The primary outcome was a composite of PPCs. The secondary outcomes were the differences in intraoperative hemodynamic values (mean arterial pressure, MAP; cardiac index, CI; systemic vascular resistance index, SVRI; and stroke volume variation, SVV).
Results
A total of 70 patients (65 men [93%]; 5 women [7%]) completed the trial; median (IQR) age was 52.0 (48-59) years in the TIVA group and 57.0 (46-64) years in the INH group. The demographic characteristics were similar between the 2 groups, except that patients in the TIVA group had a slightly lower body mass index. Patients in the TIVA group had a lower risk of developing PPCs (unadjusted odds ratio, 0.25; 95% CI, 0.08-0.80). The TIVA group had significantly higher MAP, lower CI, and higher SVRI than the INH group after the third hour of monitoring. The TIVA group showed a relatively stable hourly MAP, CI, SVRI, and SVV across time points, while the INH group showed a more varying pattern. The generalized estimating equation showed no clinical differences in the trend of hemodynamic parameters across time between groups.
Conclusions and Relevance
In this randomized clinical trial, using propofol-based TIVA reduced the incidence of PPCs in free flap surgery. This finding may be related to more stable hemodynamic manifestations and a lower total balance of fluid throughout the surgery.
Trial Registration
ClinicalTrials.gov Identifier: NCT03263078
Introduction
Microvascular free flap reconstruction is a complicated and time-consuming surgical procedure that has been performed widely since the 1970s and has become the criterion standard for the surgical treatment of head and neck cancer. Although its overall success rate is approximately 95% to 97%,1,2 the complication rate cannot be ignored. Several studies have examined the effect of propofol and inhalation anesthetics on the outcome after cancer surgery,3,4,5 but there is little evidence on their influence in free flap surgery.
Our previous retrospective study6 showed that patients who received propofol-based total intravenous anesthesia (TIVA) had lower fluid requirements and 30% fewer postoperative pulmonary complications (PPCs) during free flap surgery under a conventional fluid administration strategy. The protective effect of propofol against pulmonary complications could be explained by the milder modulatory effects on vascular resistance and attenuated inflammation from prolonged free flap surgery. The aim of this study was to examine whether this protective effect of propofol exists in a randomized setting where a goal-directed fluid strategy was applied. We hypothesized that patients in the TIVA group would have fewer postoperative complications and a more stable hemodynamic appearance than those in the inhalational anesthesia (INH) group.
Methods
Study Setting and Randomization
This study was reviewed and approved by the Taichung Veterans General Hospital Institutional Review Board (CF16113B) (protocol in Supplement 1), and written informed consent was obtained from all patients participating in the trial before receiving the operation. The trial was registered prior to patient enrollment at ClinicalTrials.gov (NCT03263078; principal investigator: Yi-Ting Chang; date of registration: August 21, 2017) and followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline. This study enrolled patients who underwent head and neck tumor resection and microvascular free flap reconstructions from October 2017 to October 2019. Patients were excluded for the following reasons: preoperative cognitive dysfunction, New York Heart Association functional classification greater than III or a left ventricular ejection fraction less than 30%; preoperative documented severe obstructive or restrictive lung disease; liver cirrhosis; severe chronic kidney insufficiency (glomerular filtration rate <30 mL/min/1.73 m2); and peripheral arterial occlusive disease, active cardiac arrhythmias, and anticipated bilateral femoral arterial and central venous catheterization when the jugular and subclavian veins were not available due to surgical extent. The last 2 criteria are known to influence the validity of hemodynamic data, which we acquired by thermodilution and arterial pulse contour analysis.
One responsible investigator who was not involved in postoperative data collection performed patient enrollment. All patients provided informed consent 1 day before the surgery. On the day of the operation, 1 investigator used computer-generated simple randomization with 1:1 ratio and randomized patients into either the INH group or the TIVA group. Nurse anesthetists received the group allocation in a sealed envelope and prepared the anesthetics. Surgeons, attending anesthesiologists, and nurse anesthetists in the operating room were not blinded to the group. Physicians who provided postoperative intensive care would know the type of anesthetics these patients had received during the surgery but did not know the hypothesis of this study.
Preoperative Management
All patients received identical preoperative management, including laboratory tests, chest radiography, electrocardiography, and medical optimization. All patients fasted from solid food 8 hours before the surgery, following the preoperative fasting guidelines. On the morning of the surgery, all patients entered the operation room after receiving 500 mL of intravenous crystalloid supplement. There was no routine premedication for anxiolysis in our institution.
Intraoperative Protocol
Anesthesia Management
Before the induction of anesthesia, routine noninvasive monitoring of blood pressure, 3-lead electrocardiography, and oxygen saturation was done. We introduced anesthesia by intravenous administration of 2 μg/kg of fentanyl, 1 mg/kg of lidocaine, and 1 to 2 mg/kgof propofol. The choice and dosage of the neuromuscular agent were based on the attending anesthesiologist’s preference. After endotracheal tube insertion, all patients underwent brachial arterial catheterization with a 4-Fr pulse contour cardiac output (PiCCO) catheter and central venous cannulation with an 8-Fr catheter (subclavian or femoral, according to surgical exposure). The PiCCO catheter line was connected to a calibrated PiCCO Module (PiCCO; PULSION Medical Systems SE) with both transpulmonary thermodilution and arterial pulse contour analysis functions.
Anesthesia was maintained using 1% to 3% sevoflurane (Ultane; AbbVie Inc) in the INH group. In the TIVA group, we used 2% propofol (Fresofol; Fresenius Kabi) administered by the target-controlled infusion system (Orchestra Base Primea; Fresenius Kabi) in Schneider mode with an effect concentration of 2.5 to 3.5 μg/mL. We applied a bispectral index sensor (BIS sensor; Covidien) for anesthesia depth monitoring and kept the BIS value at 40 to 50 in both groups. For analgesia, the attending anesthesiologists used fentanyl in both groups based on patient hemodynamic changes due to pain (tachycardia, hypertension). For sufficient neuromuscular relaxation, we infused rocuronium at 6 to 9 μg/kg/min to maintain a train-of-four (TOF) count between 1 and 2. Mechanical ventilation in both groups was set to a tidal volume of 8 mL/kg predicted body weight in pressure control mode to have a reasonable effect on stroke volume variation, positive end‐expiratory pressure was set to 5 cm H2O, and the respiratory rate was kept at 8 to 10 breaths/min to keep end-tidal carbon dioxide between 35 and 45 cm H2O.
Goal-Directed Fluid Therapy and Hemodynamic Management
In both groups, the maintenance fluid was lactated Ringer solution at a rate of 4 mL/kg/h during surgery. We used the PiCCO system to facilitate goal-directed fluid therapy intraoperatively in all patients. We set a hemodynamic goal of keeping the mean arterial pressure (MAP) at greater than 75 mm Hg or the reduction in MAP less than 10% from baseline. The rationale of this goal was to minimize the duration of MAP below 65 mm Hg, which has been associated with postoperative complications,7,8 and was also based on the cruciality of maintaining hemodynamic stability and adequate perfusion for fresh anastomosis during free flap surgery.
When patients had a MAP less than 75 mm Hg or their MAP fell by greater than 10% from baseline for more than 5 minutes, the nurse anesthetist would page the consultant anesthesiologist and initiate the hemodynamic management algorithm shown in the eFigure in Supplement 2. The framework of this algorithm was adopted according to previously published methods and results.9,10,11,12,13 In brief, fluid responsiveness was defined as stroke volume variation (SVV) at least 10% using PiCCO technology,14,15 and all patients with MAP less than 75 mm Hg or MAP more than 10% lower than baseline for more than 5 minutes were treated by 1 of 3 strategies based on the hemodynamic algorithm (intravenous fluid bolus, inotropic support, or vasopressor support).
Postoperative Management
Postoperative management was not controlled in this study, but all patients received standard intensive care unit (ICU) care according to our institutional protocols by the same team of critical care physicians. After surgery, the patients were transferred to the ICU, which specializes in postoperative care of microvascular flap reconstruction surgery. Neuromuscular blockade reversal agents were not administered to these patients because there was no immediate need for extubation in the operating room. In the ICU, these patients received sedation overnight via dexmedetomidine infusion at a dose of 0.4 to 0.7 μg/kg/h to maintain a Ramsay score of 2 without any neuromuscular blockade agents. All patients underwent at least 1 chest radiography as soon as they arrived at the ICU and underwent routine monitoring of flap perfusion to detect any possible hematoma, congestion, or bleeding. At 6 am on postoperative day 1, these patients were weaned off the ventilator according to a standard protocol once they were awake and cooperative and their muscle strength had fully recovered (defined as a TOF ratio >90%).
Outcome Measures
The primary outcome of this study was the incidence of PPCs. The definition of PPCs in this study followed the European joint taskforce published guidelines for perioperative clinical outcomes (EPCO)16 and other definitions published by Miskovic and Lumb.17 Any PPCs such as pulmonary edema (chest radiography with pulmonary congestion), pneumonia (chest radiography with at least 1 of the following: infiltrate, consolidation, cavitation, and either fever or leukocytosis), pleural effusion (chest radiography with blunting of costophrenic angle), or atelectasis (lung opacification with mediastinal shift, hilum or hemidiaphragm shift toward the affected area) reported in the postoperative phase were collected via medical record review. Surgical complications, defined as thrombosis, bleeding, and graft failure that needed a repeated operation after the initial free flap surgery, were also collected via medical record review.
The secondary outcomes of this study were the differences in the hemodynamic parameters MAP, cardiac index (CI), systemic vascular resistance index (SVRI), and SVV. These data were recorded after the start of monitoring (T0) and then hourly until the end of surgery (Tend). The frequency of algorithm initiation, types and frequencies of hemodynamic interventions required, volume of crystalloids and colloids administered, urine output, and blood loss were also recorded.
Sample Size Calculation
Based on our previous retrospective study,6 which showed a 30% difference in PPCs (TIVA vs INH: 26.1% vs 54.0%), power analysis suggested that a minimum of 64 patients (32 per group) would be required to detect significant differences in PPC rates between the INH and TIVA groups, with a type I error of .05, a power of 0.8.
Statistical Data Analyses
We performed all statistical analyses using SPSS statistical software (IBM SPSS Statistics, version 25.0). Categorical variables were compared using Pearson χ2 test, and Fisher exact test was used when the frequency was expected to be small. For continuous variables, we compared using Mann-Whitney test for non-normal distributed continuous variable and t test for normal distributed continuous variables. Clinical meaningful difference were reported with effect size and 95% CI. We used Cohen d for t test, effect size r for Mann-Whitney U test, and Φ (phi) coefficient for χ2 test. Cohen d effect size was demonstrated as small (d = 0.2), medium (d = 0.5), and large (d > 0.8); effect size r was demonstrated as small (r < 0.3), medium (0.3 < r <0.5), and large (r > 0.5); Φ coefficient effect size was demonstrated as weak (Φ = 0.1-0.2), moderate (Φ > 0.2), relatively strong (Φ > 0.4), and strong (Φ > 0.6).
Variables with potential effects on PPCs, based on a previous literature review,18,19,20 were added into the multivariate logistic regression model of PPCs using entry selection. Repeated measurements of hemodynamic data were compared within groups using the Friedman test and were compared between groups using a generalized estimating equation (GEE) to adjusted time effect and Bonferroni test for pairwise comparison.
Results
The CONSORT flow diagram can be found in Figure 1. Seventy-eight patients participated in the study. Six patients met the exclusion criteria, 1 declined to participate, and 1 had the surgery postponed. Seventy patients were randomized to the INH and TIVA groups. The adherence to the intraoperative anesthetic plan was 100%. Most of the patients were male, with American Society of Anesthesiologists (ASA) physical status classification 2, and had oral cavity tumors. A total of 16 patients (23%) had a smoking history and were diagnosed with chronic obstructive pulmonary disease. A total of 29 patients (41%) were new cases, while 30 (43%) were local recurrence cases. More than 70% (54 of 70) of the surgical procedures in this study involved bony structure excisions, such as maxillectomy or mandibulectomy. The most common reconstruction choices were anterolateral thigh (35 [50%]) and radial forearm (29 [41%]). The demographic characteristics were similar between the groups, while patients in the TIVA group had a slightly lower body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) (median [IQR], TIVA vs INH: 21.4 [18.7-25.9] vs 23.9 [21.4-26.5]; effect size r = 0.26, with small effect). The demographic characteristics data for the 2 groups are summarized in Table 1.
Figure 1. CONSORT Flow Diagram.
CONSORT flow diagram showing participant flow through each stage of the randomized clinical trial (enrollment, intervention allocation, follow-up, and data analysis). INH indicates inhalation anesthesia; TIVA, total intravenous anesthesia.
Table 1. Demographic Characteristics of the 2 Groups.
| Characteristic | No. (%) | |
|---|---|---|
| TIVA (n = 35) | INH (n = 35) | |
| Patient variables | ||
| Sex | ||
| Female | 3 (9) | 2 (6) |
| Male | 32 (91) | 33 (94) |
| Age, median (IQR), y | 52.0 (48-59) | 57.0 (46-64) |
| BMI, median (IQR) | 21.4 (18.7-25.9) | 23.9 (21.4-26.5) |
| ASA classificationa | ||
| 1 | 0 | 5 (14) |
| 2 | 26 (74) | 22 (63) |
| 3 | 9 (26) | 8 (23) |
| Hypertension | 11 (31) | 15 (43) |
| Diabetes | 5 (14) | 2 (6) |
| Chronic obstructive pulmonary disease | 9 (26) | 7 (20) |
| Mild | 7 | 6 |
| Moderate | 2 | 1 |
| Baseline MAP, median (IQR), mm Hg | 93.0 (81-104) | 88 (80-104) |
| Tumor stageb | ||
| II | 4 (11) | 9 (26) |
| III | 9 (26) | 4 (11) |
| IVA | 20 (57) | 20 (57.) |
| IVB | 2 (6) | 2 (6) |
| Tumor site | ||
| Oral cavity | 32 (91) | 28 (80) |
| Oropharynx | 2 (6) | 6 (17) |
| Hypopharynx | 0 | 1 (3) |
| Nasopharynx | 1 (3) | 0 |
| Operation indication | ||
| First time tumor excision | 12 (34) | 17 (49) |
| Local recurrence | 19 (54) | 11 (31) |
| Local recurrence and osteoradionecrosis | 1 (3) | 5 (14) |
| Wound defect | 3 (9) | 2 (6) |
| Excision type | ||
| Soft tissue only | 9 (26) | 7 (20) |
| Maxillectomy | 8 (23) | 4 (11) |
| Mandibulectomy | 18 (51) | 24 (69) |
| Reconstruction choice | ||
| Anterolateral thigh | 22 (63) | 13 (37) |
| Radial forearm | 10 (29) | 19 (54) |
| Medial sural artery perforator | 2 (6) | 1 (3) |
| Posteromedial thigh | 1 (3) | 0 |
| Fibula | 0 | 2 (6) |
Abbreviations: BMI, body mass index, calculated as weight in kilograms divided by height in meters squared; INH, inhalation group; MAP, mean arterial blood pressure; TIVA, total intravenous anesthesia.
ASA classification refers to American Society of Anesthesiologists physical status classification.
Clinical stage (American Joint Committee on Cancer, 8th edition) before the operation.
Primary Outcome
The primary outcome of this study was the incidence of PPCs. Table 2 displays the breakdown of PPCs between 2 groups. Nineteen patients developed PPCs, which made the overall PPC rate 27.1%. The TIVA group had a lower PPC incidence than the INH group (TIVA vs INH: 14.3% vs 40%; effect size Φ coefficient, 0.29; 95% CI, 0.04-0.50, with moderate effect). Logistic regression in Table 3 showed that patients in the TIVA group had a lower odds of developing PPCs compared with the INH group (absolute difference, 25.7%; unadjusted odds ratio, 0.25; 95% CI, 0.08-0.80). We included BMI in the regression model because it was the only demographic variable that showed a moderate effect size of difference between groups; we also included variables such as age and ASA classification, which have been shown in the literature to have an effect on PPCs. The multivariate logistic regression showed that using TIVA to maintain anesthesia resulted in a lower risk of PPCs than using inhalation (adjusted odds ratio, 0.26; 95% CI, 0.07-0.88).
Table 2. Intraoperative and Postoperative Data in the 2 Groups.
| No. (%) | Effect size (95% CI)a | ||
|---|---|---|---|
| TIVA (n = 35) | INH (n = 35) | ||
| Intraoperative data, median (IQR) | |||
| BISb | 45.0 (43 to 46) | 44.0 (42 to 46) | 0.10 (−0.16 to 0.31) |
| Fentanyl, μg/kg | 6.3 (4.7 to 9.0) | 4.2 (2.7 to 5.7) | 0.42 (0.22 to 0.61) |
| Propofol, mgc | 2548 (2083.2 to 3124.8) | 111 (90 to 126) | 0.86 (0.83 to 0.86) |
| Propofol, mLd | 127.4 (104.2 to 156.2) | 5.6 (4.5 to 6.3) | 0.86 (0.83 to 0.86) |
| Tracheostomy | 25 (71) | 23 (66) | 0.06 (0.00 to 0.30) |
| Anesthesia duration, he | 10.8 (10 to 11.3) | 11.3 (10.3 to 12.3) | −0.13 (−0.37 to 0.09) |
| Crystalloid, mL | 3500 (2850 to 4300) | 4300 (3400 to 5500) | −0.30 (−0.51 to −0.06) |
| Crystalloid, mL/kg/h | 5.7 (4.8 to 6.3) | 5.8 (4.9 to 7) | −0.11 (−0.35 to 0.11) |
| Colloid, mL | 0.0 (0 to 500) | 500 (500 to 1000) | −0.48 (−0.65 to −0.29) |
| Colloid, mL/kg/h | 0.0 (0 to 0.7) | 0.8 (0.5 to 1.1) | −0.41 (−0.61 to −0.18) |
| Blood loss, mL | 710 (450 to 1030) | 650 (450 to 905) | 0.04 (−0.21 to 0.26) |
| Urine output, mL | 1250 (800 to 1900) | 1250 (950 to 2300) | −0.11 (−0.34 to 0.11) |
| Fluid balance, mLf | 2113 (1118 to 2739) | 2810 (1934 to 3458) | −0.32 (−0.53 to −0.09) |
| Patients need interventiong | 7 (20) | 23 (66) | 0.46 (0.23 to 0.67) |
| Hemodynamic intervention, mean (SD), timesh | |||
| Volume bolus | 1.3 (0.5) | 1.7 (1.8) | −0.3 (−0.95 to 0) |
| Inotropic | 0.1 (0.4) | 0.4 (0.8) | −0.5 (−2.32 to −1.22) |
| Vasopressor | 0.0 (0.0) | 1.0 (0.8) | −1.8 (−1.31 to −0.34) |
| Overall intervention | 1.4 (0.5) | 3.0 (2.7) | −0.8 (−0.77 to 0.17) |
| Postoperative data | |||
| Pulmonary complication | 5 (14.3) | 14 (40.0) | 0.29 (0.04 to 0.50) |
| Pulmonary edema | 0 | 5 (35.7) | NA |
| Pneumonia | 1 (20.0) | 4 (28.6) | NA |
| Pleural effusion | 3 (60.0) | 3 (21.4) | NA |
| Atelectasis | 0 | 2 (14.3) | NA |
| Pneumothorax | 1 (20.0) | 0 | NA |
| Surgical complication | 8 (22.9) | 9 (25.7) | 0.03 (0.00 to 0.28) |
| Surgical site infection | 6 (75.0) | 4 (44.4) | NA |
| Flap congestion/failure | 2 (25.0) | 1 (11.1) | NA |
| Bleeding hematoma | 0 | 4 (44.4) | NA |
| Mechanical ventilation, d | 1.5 (0.8 to 1.8) | 1.6 (0.9 to 2.5) | −0.21 (−0.42 to 0.02) |
| ICU stay, d | 5.8 (4.6 to 6) | 5.8 (4.8 to 6.7) | −0.05 (−0.28 to 0.19) |
| Postoperative hospital stay, d | 16.0 (14 to 22) | 15.0 (14 to 18) | 0.15 (−0.08 to 0.38) |
Abbreviations: BIS, bispectral index; ICU, intensive care unit; INH, inhalation; NA, not applicable; TIVA, total intravenous anesthesia.
Effect size presented as Cohen d or φ coefficient value (95% CI of effect size).
Bispectral index, an electroencephalogram-based anesthesia depth monitor; BIS values between 40 and 60 represent adequate general anesthesia for a surgery.
Propofol (mg), the total propofol dosage in the 2 groups. Patients in both groups received propofol for induction of general anesthesia, and the INH group had inhalation anesthetic while the TIVA group had propofol as maintenance of general anesthesia.
Propofol (mL) equals to propofol (mg)/20, as we use 2% propofol in this study.
Anesthesia duration from the induction of anesthesia to the end of the surgery.
Fluid balance defined as (propofol volume [mL] + crystalloid + colloid) − (blood loss + urine output).
Patients who need hemodynamic intervention are defined as the number of patients who had at least 1 episode of mean arterial pressure less than 75 mm Hg or a reduction of 10% from baseline.
Data here are the mean values of the intervention frequency (times) among patients who need hemodynamic intervention during the surgery.
Table 3. Logistic Regression of Postoperative Pulmonary Complicationsa.
| OR (95% CI) | ||
|---|---|---|
| Unadjusted | Adjusted | |
| TIVA vs INH | 0.25 (0.08-0.80) | 0.26 (0.07-0.88) |
| Age | 1.07 (1.00-1.14) | 1.05 (0.98-1.12) |
| BMI | 1.07 (0.93-1.23) | 1.04 (0.89-1.22) |
| ASA 3 vs 1 and 2 | 2.39 (0.75-7.63) | 2.03 (0.53-7.79) |
Abbreviations: ASA, American Society of Anesthesiologists physical status classification; BMI, body mass index; INH, inhalation; OR, odds ratio; TIVA, total intravenous anesthesia.
Logistic regression. Patients in the TIVA group had a lower incidence of postoperative pulmonary complications. We adjusted the logistic regression of postoperative pulmonary complications (entry method) with variables that had differences between groups (BMI) or had an effect on postoperative complications according to the literature reviews (age, ASA class 3).
Secondary Outcomes
We recorded the MAP, CI, SVRI, and SVV values from the pulse contour analysis of the PiCCO system hourly and compared these parameters within groups using the Friedman test. The TIVA group showed a relatively stable pattern of hourly MAP, CI, SVRI, and SVV values across time while the INH had a more varying pattern across time. The eTable in Supplement 2 summarizes the difference of MAP, CI, SVRI, and SVV within groups and between groups.
Patients in the TIVA group had higher MAP than those in the INH group from T2 to Tend with medium effect size (TIVA vs INH, median [IQR] mm Hg: T2, 95 [86-103] vs 84 [75-94]; effect size r, 0.38; 95% CI, 0.16-0.58; T3, 96 [86-105] vs 82 [74-91]; effect size r, 0.47; 95% CI, 0.27-0.65; T4, 93 [87-102] vs 82 [75-88]; effect size r, 0.54; 95% CI, 0.35-0.71]; T5, 92 [86-99] vs 76 [73-80]; effect size r, 0.58; 95% CI, 0.40-0.74; T6, 94 [83-98] vs 79 [74-85]; effect size r, 0.47; 95% CI, 0.26-0.65; Tend, 94 [86-102] vs 83 [77-88]; effect size r, 0.44; 95% CI, 0.23-0.64). The TIVA group had lower CI compared with the INH group with medium effect size at T5 and Tend. The TIVA group had a higher SVRI compared with the INH group with medium effect size at T2, T3, T4, T5, T6, and Tend. Detailed data on intraoperative hemodynamic parameters are listed in the eTable in Supplement 2. Figure 2 summarizes the comparison of hemodynamic trend between groups throughout the surgery and the mean difference at different time points after adjusting the time effect. Generalized estimating equation revealed that the trends of hourly MAP, CI, SVRI, or SVV values across time between groups were comparable (GEE, P values of MAP, CI, SVRI, and SVV = .42, .98, .56, .36, respectively). Bonferroni test for pairwise comparison showed that patients in the TIVA group had clinical meaningful higher MAP at T3, T4, T6, and Tend and higher SVRI at T3, T4, T5, T6, and Tend.
Figure 2. Changes in Hemodynamic Parameters Between the TIVA and INH Groups Over Time.
Number at risk: 35 in each group throughout T0 to Tend. P values less than .05 represented significant difference between groups across time by generalized estimating equation. The mean difference (TIVA minus INH), effect size (Cohen d), and 95% CI (Cohen d) at each time point are illustrated at the bottom of each panel. Although there were no significant differences in the trend of (A) MAP, (B) CI, (C) SVRI, or (D) SVV between groups across time, patients in the TIVA group had clinical meaningful higher MAP at T3, T4, T6, and Tend and higher SVRI at T3, T4, T5, T6, and Tend after adjusted time effect. CI indicates cardiac index; INH, inhalation anesthesia; MAP, mean arterial blood pressure; PiCCO, pulse contour cardiac output; SVRI, systemic vascular resistance index; SVV, stroke volume variation; TIVA, total intravenous anesthesia; T0: data after instrumentation of PiCCO; T1: first hour; T2: 2nd hour; Tend: the end of surgery.
aP value less than .05 in pairwise comparison by Bonferroni test.
Intraoperative Data in the 2 Groups
The intraoperative variables are listed in Table 2. Patients in the TIVA group had a significantly higher intraoperative fentanyl requirement than patients in the INH group (TIVA vs INH: median [IQR], 6.3 [4.7-9.0] μg/kg vs 4.2 [2.7-5.7] μg/kg; effect size r, 0.42; 95% CI, 0.22-0.61), while the mean BIS value was similar between groups. Only a subset of patients needed hemodynamic intervention throughout the surgery. A total of 7 patients (20%) in the TIVA group needed hemodynamic intervention throughout the surgery, while 23 patients (66%) in the INH group needed intervention (TIVA vs INH: effect size Φ coefficient, 0.46; 95% CI, 0.23-0.67). Patients in the TIVA group had lower crystalloid and colloid requirements than patients in the INH group with medium effect size (TIVA vs INH: crystalloid, 3500 mL vs 4300 mL; effect size r, −0.30; 95% CI, −0.51 to −0.06); colloid, 0 mL vs 500 mL; effect size r, −0.48; 95% CI, −0.65 to −0.29). Because patients in the TIVA group received a certain amount of intravenous propofol, the volume of intravenous propofol was also included in the input volume when calculating the overall fluid balance. Overall fluid balance was also lower in the TIVA group with medium effect size (TIVA vs INH: 2113 mL vs 2810 mL; effect size r, −0.32; 95% CI, −0.53 to −0.09), while urine output and surgical blood loss were similar between groups.
Postoperative Data in the 2 Groups
The postoperative data are summarized in Table 2. One patient in the INH group died from intractable carotid artery bleeding during hospitalization. The overall in-hospital mortality was 1.4% (1 of 70 patients). The overall surgical complication rate was 24.2% (17 of 70 patients). There was no difference in surgical complications between the TIVA and INH groups. The duration of postoperative mechanical ventilation, length of stay in the ICU, and length of stay in hospital were not different between 2 groups.
Discussion
This randomized clinical trial examined 70 patients who were randomized to the TIVA or INH group. The primary outcome was the incidence of PPCs. The TIVA group had clinically meaningful fewer PPCs than the INH group, and the difference remained significant after multivariate logistic regression. The TIVA group had fewer patients who needed hemodynamic intervention throughout the free flap surgery and showed a relatively stable pattern of hourly MAP, CI, SVRI, and SVV across time.
Patients with head and neck cancer who received tumor excision and free flap reconstruction are at high risk of developing PPCs. In the current study, nearly 1 of 3 patients developed PPCs, which was a relatively high incidence compared with other major studies in the last decades that reported the incidence of PPCs to range from 1% to 23%.21,22,23,24,25,26 This may be related to the fact that oral cavity cancer was the dominant tumor site in the current study. Both TIVA and INH groups had more than 80% of patients with oral cavity cancer, and it can be presumed that some of them would be affected by subtle aspiration in the postoperative period, which may make a contribution to the high PPC incidence.
The most important finding of the current study is that patients in the TIVA group had a clinically meaningful lower PPC rate, and this reduction remained significant after adjusting for age, ASA classification, and BMI in multivariate logistic regression. This can be explained by 2 facts. First, although both groups in the current study had the same maintenance fluid rate (4 mL/kg/h), the TIVA group had a lower intraoperative fluid requirement and less overall fluid balance, even though the volume of intravenous propofol had been considered in the calculation of the sum of fluid balance. This can account for fewer patients in the TIVA group having MAP drops below the threshold that would initiate the hemodynamic intervention algorithm. The first step of the algorithm in the current study was to optimize the SVV. These patients with fluid responsiveness would initially receive crystalloid, 250 mL. If the MAP did not increase, they received another 250-mL crystalloid bolus and an extra 250 mL of colloid. Therefore, the more the hemodynamic algorithm was initiated, more intravenous fluids were likely to be provided, and patients with MAP drops below the threshold more frequently were more likely to be hypervolemic. Hypervolemia can induce the release of atrial natriuretic peptide and dilution of plasma protein, both of which can deteriorate the endothelial glycocalyx.27,28,29 Losing the integrity of the endothelial glycocalyx in the pulmonary capillary system can start a vicious circle of increased vascular permeability, fluid extravasation, inflammation, and induced PPCs.
Second, the rarer PPCs in the TIVA group could be explained by the protective effect of propofol. Propofol has been reported to modulate lipopolysaccharide-induced acute lung injury, downregulate HMGB1 expression in human alveolar epithelial cells, and suppress pyroptosis.30,31,32 Free flap surgery is a prolonged surgery that carries the risk of ischemia and reperfusion injury during the period of donor flap harvesting and anastomosis. Therefore, we consider propofol to have a protective effect against PPCs by maintaining the stability of the glycocalyx owing to a lower intravenous fluid requirement and the modulation of the systemic inflammatory response.
In this study, the trend of each hemodynamic parameter throughout the surgery was similar between groups. It was logical that both groups shared trends of decreased MAP, increased CI, and decreased SVRI that were associated with cardiovascular suppression from anesthetics, but the interesting finding was that the TIVA group had higher MAP and higher SVRI starting from the third hour of monitoring. This cannot be explained by inadequate anesthesia or analgesia in the TIVA group. The BIS value was comparable between groups. The TIVA group had a higher intraoperative fentanyl requirement because of propofol’s hypnotic effect through potentiation of the effects of the inhibitory neurotransmitter GABA but has long been considered nonanalgesic33 or to provide only mild analgesia at subhypnotic doses.34,35,36 Another explanation would be the modulation of vascular resistance from propofol. Sudheer et al37 found relative cardiovascular stability in their propofol group because it had significantly higher SVRI than the inhalation group when changing the patients’ position. Deryck et al38 demonstrated that propofol did not affect the left ventricle afterload or ventricular performance, while sevoflurane decreased arterial vessel tone and reduced left ventricular-arterial coupling. Another study39 demonstrated that patients in the volatile group had low SVR and a high CI and required more vasopressors in the immediate postcardiopulmonary bypass period of coronary artery bypass graft. Therefore, we postulate that the milder modulation of vascular resistance is what led to the more stable hemodynamic properties in the TIVA group.
Another interesting finding is that the lower PPC rate in the TIVA group did not shorten the length of hospitalization and ICU stay. It may be related to early diagnosis and intervention, and most of the PPCs were mild without causing respiratory failure. Nevertheless, the average length of stay in the hospital and in the ICU in this study was relatively prolonged compared with other published reviews.21,22,23,24,25,26 Patients in our country had national health insurance to cover most of the medical expenses; therefore, patients were in favor of remaining hospitalized until they had wound stitches removed, finished the rehabilitation program, and corked the tracheostomy successfully. Some patients met the discharge criteria of the ICU but delayed the transfer because they were waiting for a single-room ward.
Limitations
The main limitation of this work is that goal-directed fluid management was not done in the postoperative period. In our institution, intensivists care for these patients without following the intraoperative program, and they may administer fluids based on their clinical experience. As some of the patients in this study were receiving ventilator support more than 24 hours after surgery, the postoperative fluid management could interfere with the result of PPCs. The second limitation is the lack of data on functional recovery and the serum cytokine profile, which would have helped us further evaluate the differences between the 2 anesthetics. The third limitation is the relatively small sample size of this study, as the sample size calculation was based on PPC data from our previous retrospective study. Larger, multicenter randomized clinical trials are needed to confirm the effects seen in this study and increase the patient population to increase the generalizability of results.
Conclusions
In this randomized clinical trial, patients undergoing free flap surgery who received TIVA had a lower incidence of PPCs than patients who received INH. This finding may be related to their more stable hemodynamic manifestations throughout the surgery and to their lower total balance of fluid.
Trial Protocol
eFigure. Hemodynamic assessment and intervention algorithm
eTable. Intraoperative hemodynamic parameters in the two groups
Data Sharing Statement
References
- 1.Haughey BH, Wilson E, Kluwe L, et al. Free flap reconstruction of the head and neck: analysis of 241 cases. Otolaryngol Head Neck Surg. 2001;125(1):10-17. doi: 10.1067/mhn.2001.116788 [DOI] [PubMed] [Google Scholar]
- 2.Pohlenz P, Blessmann M, Heiland M, Blake F, Schmelzle R, Li L. Postoperative complications in 202 cases of microvascular head and neck reconstruction. J Craniomaxillofac Surg. 2007;35(6-7):311-315. doi: 10.1016/j.jcms.2007.05.001 [DOI] [PubMed] [Google Scholar]
- 3.Sessler DI, Riedel B. Anesthesia and cancer recurrence: context for divergent study outcomes. Anesthesiology. 2019;130(1):3-5. doi: 10.1097/ALN.0000000000002506 [DOI] [PubMed] [Google Scholar]
- 4.Yap A, Lopez-Olivo MA, Dubowitz J, Hiller J, Riedel B; Global Onco-Anesthesia Research Collaboration Group . Anesthetic technique and cancer outcomes: a meta-analysis of total intravenous versus volatile anesthesia. Can J Anaesth. 2019;66(5):546-561. doi: 10.1007/s12630-019-01330-x [DOI] [PubMed] [Google Scholar]
- 5.Hong B, Lee S, Kim Y, et al. Anesthetics and long-term survival after cancer surgery-total intravenous versus volatile anesthesia: a retrospective study. BMC Anesthesiol. 2019;19(1):233. doi: 10.1186/s12871-019-0914-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Chang YT, Wu CC, Tang TY, Lu CT, Lai CS, Shen CH. Differences between total intravenous anesthesia and inhalation anesthesia in free flap surgery of head and neck cancer. PLoS One. 2016;11(2):e0147713. doi: 10.1371/journal.pone.0147713 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Salmasi V, Maheshwari K, Yang D, et al. Relationship between intraoperative hypotension, defined by either reduction from baseline or absolute thresholds, and acute kidney and myocardial injury after noncardiac surgery: a retrospective cohort analysis. Anesthesiology. 2017;126(1):47-65. doi: 10.1097/ALN.0000000000001432 [DOI] [PubMed] [Google Scholar]
- 8.Sessler DI, Sigl JC, Kelley SD, et al. Hospital stay and mortality are increased in patients having a “triple low” of low blood pressure, low bispectral index, and low minimum alveolar concentration of volatile anesthesia. Anesthesiology. 2012;116(6):1195-1203. doi: 10.1097/ALN.0b013e31825683dc [DOI] [PubMed] [Google Scholar]
- 9.Kim HJ, Kim EJ, Lee HJ, et al. Effect of goal-directed haemodynamic therapy in free flap reconstruction for head and neck cancer. Acta Anaesthesiol Scand. 2018;62(7):903-914. doi: 10.1111/aas.13100 [DOI] [PubMed] [Google Scholar]
- 10.Derichard A, Robin E, Tavernier B, et al. Automated pulse pressure and stroke volume variations from radial artery: evaluation during major abdominal surgery. Br J Anaesth. 2009;103(5):678-684. doi: 10.1093/bja/aep267 [DOI] [PubMed] [Google Scholar]
- 11.Marik PE, Cavallazzi R, Vasu T, Hirani A. Dynamic changes in arterial waveform derived variables and fluid responsiveness in mechanically ventilated patients: a systematic review of the literature. Crit Care Med. 2009;37(9):2642-2647. doi: 10.1097/CCM.0b013e3181a590da [DOI] [PubMed] [Google Scholar]
- 12.Ramsingh DS, Sanghvi C, Gamboa J, Cannesson M, Applegate RL II. Outcome impact of goal directed fluid therapy during high risk abdominal surgery in low to moderate risk patients: a randomized controlled trial. J Clin Monit Comput. 2013;27(3):249-257. doi: 10.1007/s10877-012-9422-5 [DOI] [PubMed] [Google Scholar]
- 13.Pestaña D, Espinosa E, Eden A, et al. Perioperative goal-directed hemodynamic optimization using noninvasive cardiac output monitoring in major abdominal surgery: a prospective, randomized, multicenter, pragmatic trial: POEMAS Study (PeriOperative goal-directed thErapy in Major Abdominal Surgery). Anesth Analg. 2014;119(3):579-587. doi: 10.1213/ANE.0000000000000295 [DOI] [PubMed] [Google Scholar]
- 14.Berkenstadt H, Margalit N, Hadani M, et al. Stroke volume variation as a predictor of fluid responsiveness in patients undergoing brain surgery. Anesth Analg. 2001;92(4):984-989. doi: 10.1097/00000539-200104000-00034 [DOI] [PubMed] [Google Scholar]
- 15.Reuter DA, Kirchner A, Felbinger TW, et al. Usefulness of left ventricular stroke volume variation to assess fluid responsiveness in patients with reduced cardiac function. Crit Care Med. 2003;31(5):1399-1404. doi: 10.1097/01.CCM.0000059442.37548.E1 [DOI] [PubMed] [Google Scholar]
- 16.Jammer I, Wickboldt N, Sander M, et al. ; European Society of Anaesthesiology (ESA) and the European Society of Intensive Care Medicine (ESICM); European Society of Anaesthesiology; European Society of Intensive Care Medicine . 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. 2015;32(2):88-105. doi: 10.1097/EJA.0000000000000118 [DOI] [PubMed] [Google Scholar]
- 17.Miskovic A, Lumb AB. Postoperative pulmonary complications. Br J Anaesth. 2017;118(3):317-334. doi: 10.1093/bja/aex002 [DOI] [PubMed] [Google Scholar]
- 18.le Nobel GJ, Higgins KM, Enepekides DJ. Predictors of complications of free flap reconstruction in head and neck surgery: analysis of 304 free flap reconstruction procedures. Laryngoscope. 2012;122(5):1014-1019. doi: 10.1002/lary.22454 [DOI] [PubMed] [Google Scholar]
- 19.Patel RS, McCluskey SA, Goldstein DP, et al. Clinicopathologic and therapeutic risk factors for perioperative complications and prolonged hospital stay in free flap reconstruction of the head and neck. Head Neck. 2010;32(10):1345-1353. doi: 10.1002/hed.21331 [DOI] [PubMed] [Google Scholar]
- 20.Petrar S, Bartlett C, Hart RD, MacDougall P. Pulmonary complications after major head and neck surgery: a retrospective cohort study. Laryngoscope. 2012;122(5):1057-1061. doi: 10.1002/lary.23228 [DOI] [PubMed] [Google Scholar]
- 21.Yang CK, Teng A, Lee DY, Rose K. Pulmonary complications after major abdominal surgery: National Surgical Quality Improvement Program analysis. J Surg Res. 2015;198(2):441-449. doi: 10.1016/j.jss.2015.03.028 [DOI] [PubMed] [Google Scholar]
- 22.Fisher BW, Majumdar SR, McAlister FA. Predicting pulmonary complications after nonthoracic surgery: a systematic review of blinded studies. Am J Med. 2002;112(3):219-225. doi: 10.1016/S0002-9343(01)01082-8 [DOI] [PubMed] [Google Scholar]
- 23.Smith PR, Baig MA, Brito V, Bader F, Bergman MI, Alfonso A. Postoperative pulmonary complications after laparotomy. Respiration. 2010;80(4):269-274. doi: 10.1159/000253881 [DOI] [PubMed] [Google Scholar]
- 24.McAlister FA, Bertsch K, Man J, Bradley J, Jacka M. Incidence of and risk factors for pulmonary complications after nonthoracic surgery. Am J Respir Crit Care Med. 2005;171(5):514-517. doi: 10.1164/rccm.200408-1069OC [DOI] [PubMed] [Google Scholar]
- 25.Canet J, Sabaté S, Mazo V, et al. ; PERISCOPE group . Development and validation of a score to predict postoperative respiratory failure in a multicentre European cohort: a prospective, observational study. Eur J Anaesthesiol. 2015;32(7):458-470. doi: 10.1097/EJA.0000000000000223 [DOI] [PubMed] [Google Scholar]
- 26.Kor DJ, Warner DO, Alsara A, et al. Derivation and diagnostic accuracy of the surgical lung injury prediction model. Anesthesiology. 2011;115(1):117-128. doi: 10.1097/ALN.0b013e31821b5839 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Chappell D, Bruegger D, Potzel J, et al. Hypervolemia increases release of atrial natriuretic peptide and shedding of the endothelial glycocalyx. Crit Care. 2014;18(5):538. doi: 10.1186/s13054-014-0538-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Şentürk M, Orhan Sungur M, Sungur Z. Fluid management in thoracic anesthesia. Minerva Anestesiol. 2017;83(6):652-659. doi: 10.23736/S0375-9393.17.11760-8 [DOI] [PubMed] [Google Scholar]
- 29.Brettner F, von Dossow V, Chappell D. The endothelial glycocalyx and perioperative lung injury. Curr Opin Anaesthesiol. 2017;30(1):36-41. doi: 10.1097/ACO.0000000000000434 [DOI] [PubMed] [Google Scholar]
- 30.Wang X, Liu C, Wang G. Propofol protects rats and human alveolar epithelial cells against lipopolysaccharide-induced acute lung injury via inhibiting HMGB1 expression. Inflammation. 2016;39(3):1004-1016. doi: 10.1007/s10753-016-0330-6 [DOI] [PubMed] [Google Scholar]
- 31.Zhou CH, Zhu YZ, Zhao PP, et al. Propofol inhibits lipopolysaccharide-induced inflammatory responses in spinal astrocytes via the toll-like receptor 4/MyD88-dependent nuclear factor-κB, extracellular signal-regulated protein kinases1/2, and p38 mitogen-activated protein kinase pathways. Anesth Analg. 2015;120(6):1361-1368. doi: 10.1213/ANE.0000000000000645 [DOI] [PubMed] [Google Scholar]
- 32.Liu Z, Meng Y, Miao Y, Yu L, Yu Q. Propofol reduces renal ischemia/reperfusion-induced acute lung injury by stimulating sirtuin 1 and inhibiting pyroptosis. Aging (Albany NY). 2020;13(1):865-876. doi: 10.18632/aging.202191 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Grounds RM, Lalor JM, Lumley J, Royston D, Morgan M. Propofol infusion for sedation in the intensive care unit: preliminary report. Br Med J (Clin Res Ed). 1987;294(6569):397-400. doi: 10.1136/bmj.294.6569.397 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Bandschapp O, Filitz J, Ihmsen H, et al. Analgesic and antihyperalgesic properties of propofol in a human pain model. Anesthesiology. 2010;113(2):421-428. doi: 10.1097/ALN.0b013e3181e33ac8 [DOI] [PubMed] [Google Scholar]
- 35.Anker-Møller E, Spangsberg N, Arendt-Nielsen L, Schultz P, Kristensen MS, Bjerring P. Subhypnotic doses of thiopentone and propofol cause analgesia to experimentally induced acute pain. Br J Anaesth. 1991;66(2):185-188. doi: 10.1093/bja/66.2.185 [DOI] [PubMed] [Google Scholar]
- 36.Sahinovic MM, Struys MMRF, Absalom AR. Clinical pharmacokinetics and pharmacodynamics of propofol. Clin Pharmacokinet. 2018;57(12):1539-1558. doi: 10.1007/s40262-018-0672-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Sudheer PS, Logan SW, Ateleanu B, Hall JE. Haemodynamic effects of the prone position: a comparison of propofol total intravenous and inhalation anaesthesia. Anaesthesia. 2006;61(2):138-141. doi: 10.1111/j.1365-2044.2005.04464.x [DOI] [PubMed] [Google Scholar]
- 38.Deryck YL, Fonck K, DE Baerdemaeker L, Naeije R, Brimioulle S. Differential effects of sevoflurane and propofol anesthesia on left ventricular-arterial coupling in dogs. Acta Anaesthesiol Scand. 2010;54(8):979-986. doi: 10.1111/j.1399-6576.2010.02267.x [DOI] [PubMed] [Google Scholar]
- 39.Soliz JM, Ifeanyi IC, Katz MH, et al. Comparing postoperative complications and inflammatory markers using total intravenous anesthesia versus volatile gas anesthesia for pancreatic cancer surgery. Anesth Pain Med. 2017;7(4):e13879. doi: 10.5812/aapm.13879 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial Protocol
eFigure. Hemodynamic assessment and intervention algorithm
eTable. Intraoperative hemodynamic parameters in the two groups
Data Sharing Statement