Assessment of Fluid Responsiveness After Tidal Volume Challenge During Pressure-Controlled Ventilation Volume Guaranteed: An observational study

Background: The reliability of pulse pressure variation (PPV) and stroke volume variation (SVV) to predict uid responsiveness have not previously been established when using pressure-controlled ventilation-volume guaranteed (PCV-VG) mode. We hypothesized that with a transient increase in tidal volume from 6 to 8 mL/kg of predicted body weight (PBW), which we reference as the “tidal volume challenge (TVC)”, the changes to PPV and SVV will be an indicator of uid responsiveness. Methods: The patients were rst ventilated with a tidal volume of (V t ) 6 mL/kg of predicted body weight (PBW) using PCV-VG. Following intravenous anesthesia induction, PPV 6 and SVV 6 were recorded, then the TVC was performed, which increased V t from 6 mL/kg to 8 mL/kg PBW for 1 minute and PPV 8 and SVV 8 were recorded again. The changes in value of PPV and SVV (ΔPPV 6-8 and ΔSVV 6-8 ) were calculated after TVC. Following the minute of TVC, the tidal volume was returned to 6 ml/kg PBW for the uid challenge (FC), a colloid infusion of 6ml/kg PBW for 20 minutes. Patients were classied as responders if there was an increase in cardiac index (CI) of more than 15% after FC, otherwise the patients were identied as non-responders. Eligible patients were divided into groups of responders or non-responders. Results: 37 patients were classied as responders and 44 were non-responders. PPV 6 and SVV 6 could not predict the uid responsiveness, while PPV 8 and SVV 8 could predict the uid responsiveness when using PCV-VG mode. The changes in value of PPV and SVV after TVC (ΔPPV 6-8 and ΔSVV 6-8 ) identied true uid responders with the highest sensitivity and specicity in the above variables, which predicted uid responsiveness with the area under the receiver operating characteristic curves (AUCs) (95% CIs) being 0.96 (0.93-1.00) and 0.98 (0.96-1.00), respectively. No signicant difference was found when comparing the AUCs of ΔPPV 6-8 and ΔSVV 6-8 (P > 0.05). Linear correlation was represented between the change value of CI after FC and the change value of SVV or PPV after TVC (r = 0.68; P < 0.0001 and r = 0.77; P < 0.0001, respectively). Conclusions: A transient increase in tidal volume, which we reference as the “tidal volume challenge (TVC)” could enhance the predictive value of PPV and SVV for the evaluation of uid responsiveness in patients under ventilation with PCV-VG.


Background
Fluid therapy is cornerstone of the hemodynamic management of patients undergoing major surgery. Goaldirected uid therapy (GDFT) attempts to avoid the occurrence of excessive uid load and hypovolemia, shorten the length of hospital stays, reduce mortality and minimize the complications caused by uid imbalance during surgery [1,2].To better implement GDFT, we need to nd appropriate indicators that can accurately predict uid responsiveness.
Studies have shown that dynamic indicators, such as pulse pressure variability (PPV) and stroke volume variability (SVV), are superior to static indices to predict uid responsiveness [2][3][4]. However, accuracy and thresholds are affected by many factors, such as changes in the mode of ventilation or tidal volume [5,6]. In surgeries involving general anesthesia, the application of volume-controlled ventilation (VCV) when attempting to ensure adequate respiration often leads to excessive airway pressure. Pressure-controlled ventilation (PCV) often fails to provide adequate ventilation volume under the condition of ensuring proper airway pressure [7,8].
Recently, a new type of intelligent ventilation mode, PCV-VG, has emerged, it combines the advantages of both VCV and PCV which utilizes a decelerating ow and constant pressure to offer the pre-set V t without increasing airway pressures [9]. In addition, application of PCV-VG combined with low tide volume (V t ) (6-8 ml/kg PBW) had been advocated as a lung protective ventilation strategy, which can effectively reduce airway pressure, avoid the occurrence of pulmonary barotrauma, stabilize hemodynamics and reduce pulmonary complications [10,11].
When the gas was delivered with decelerating ow rate in the ventilation with PCV-VG, the variability of intrathoracic pressure may decrease, which affects the accuracy of the prediction produced for uid responsiveness [8,12]. Additionally, previous studies reported that ventilation with the V t below 8 ml/kg PBW is not enough to cause signi cant changes in intrathoracic pressure, and this may blunt the accuracy of PPV and SVV to predict the uid responsiveness [13,14]. Therefore, it is not clear whether PPV and SVV could accurately predict patients' uid responsiveness during ventilation with PCV-VG combined with a lung protective ventilation strategy achieved by low V t (6-8 ml/kg PBW).
We have failed to identify any previous studies exploring uid responsiveness during ventilation with PCV-VG.
This study was conducted to assess the reliability of PPV and SVV in predicting uid responsiveness of patients ventilated using PCV-VG with low V t (6-8 ml/kg PBW) and also to determine whether the TVC could enhance the predictive value of PPV and SVV in this lung protective ventilation strategy.

General information
This prospective study was conducted between January 2020 and March 2020 in accordance with the Declaration of Helsinki principles. The study was registered in the Chinese Clinical Trial Registry (ChiCTR2000028995) and was approved by the Ethics Committee for Clinical Trials of the Second Hospital of Anhui Medical University, Hefei, China [approval no: PJ-YX2019-037]. Written informed consent was obtained from the participants or guardians before surgery. Included in this study were eighty-one patients with American Society of Anesthesiologists (ASA) physical status of I-II, aged 18 to 65 years, scheduled for laparoscopy-assisted radical gastrectomy under general anesthesia. Patients were excluded according to the following criteria: kidney dysfunction; cardiac arrhythmias and valvular heart disease; chronic obstructive pulmonary disease; right ventricular failure; intracranial hypertension; severe obesity di culty (BMI > 30 kg/m 2 ); airway asthma or a long history of smoking. FC was performed to identify uid responders if there was an increase in CI ≥ 15%, otherwise the patients were identi ed as non-responders. The patients were separated into two groups, responder and nonresponder.

Perioperative management
Intravenous and arterial access were established for monitoring of continuous invasive blood pressure and central venous pressure (CVP) after arrival to the operating room and before the induction of anesthesia. Drager monitor (model: A7, Hefei unity medical co.) was used for monitoring the non-invasive blood pressure, invasive radial arterial blood pressure, pulse oximetry, heart rate (HR), ratio of the HR and respiratory rate (HR/RR), electrocardiogram, peak airway pressure, plateau pressure (P pl ), driving pressure (P pl -positive end-expiratory pressure [PEEP]) and partial pressure of carbon dioxide in end expiratory gas (P ET CO2). The vital variables, such as pulse pressure variability (PPV); stroke volume variability (SVV) and cardiac index (CI) were recorded by a continuous non-invasive arterial pressure monitor (CNAP) (model: CANP™ Monitor 500, Guangzhou Xinju Science and Trade Co.). All patients received uid of Ringer's solution at a rate of 5 ~ 7 ml/kg/h. And bispectral index (BIS) (the America, Covidien IIc Co.) and the arterial blood gas analysisis of patients were monitored perioperatively.
General anesthesia was induced by intravenous (i.v.) midazolam (0.03 mg/kg), sufentanil (0.5 µg/kg), rocuronium (0.6 mg/kg) and etomidate (0.2 mg/kg) after mask oxygen inhalation for two minutes. Tracheal intubation was performed after reaching BIS between 40 and 60 and appropriate muscle relaxation was achieved, the Mindray anesthesia machine (model: A5, Shenzhen Mindray biomedical electronics co LTD) was connected for mechanical ventilation with PCV-VG mode and followed the setting: V t of 6 ml/kg PBW; P ET CO 2 remained in the 35-45 mmHg range with an inspired oxygen fraction 0.5 and a fresh gas ow 2 L/min of oxygen and air; PEEP was kept between 3-5 cmH 2 O and SpO 2 was maintained at more than 95%. Maintenance of anesthesia was conducted by propofol (4 ~ 8 mg/kg/h), remifentanil (0.1 ~ 0.3 µg/kg/min), cisatracurium (0.1 ~ 0.2 mg/kg/h) and sevo urane (1%~2%). Anesthetic depth was maintained at a BIS of 40-60 throughout the surgical procedure by adjusting the end-tidal concentration of sevo urane. Simultaneously, the appropriate application of vasoactive drugs (atropine and noradrenaline) was for the maintenance of HR and mean arterial pressure (MAP) at around 20% of the base value.

Hemodynamic and Respiratory monitoring and Study Protocol
The PPV, SVV, MAP, HR, CVP, peak airway pressure, P pl , driving pressure, central venous oxygen saturation, oxygenation index and compliance of the respiratory system (C rs ) were recorded after intravenous anesthesia induction. The depth of the anesthesia was ensured by the monitoring of BIS. Before the start of the surgery and administration of any vasopressor, with the patient supine, the study protocol ( Fig. 1) was started as follows: (1) Patients were ventilated by PCV-VG with V t 6 mL/kg PBW, maximum airway pressure of 30 mmH 2 O, and PEEP 3 5 cmH 2 O for 1 minute, baseline measurements including the PPV (PPV 6 ) and SVV (SVV 6 ) were recorded (T 0 ).

Sample size and statistical analyses
TVC has not previously been studied in patients ventilated with the ventilation of PCV-VG mode, the expected areas under the curves of PPV 6 and ΔPPV 6 − 8 were respectively 0.65 and 0.90, which was used to calculate the sample size requirement for comparing two ROC curves by statstodo, a statistical analytics tool. A sample size of Page 5/21 60 patients, was su cient to detect a signi cant difference (α = 5%) with a statistical power (β-value) of 90%, however it was increased to 81 patients for considering a 15% failure rate.
The continuous variables were presented as mean (standard deviation) or median (interquartile range) depending on the normal distribution tested by the D'Agostino-Pearson. Categorical variables presented as proportions (percentage), such as ASA physical status and gender, were analyzed using the χ 2 test. Continuous variables normally distributed were analyzed using independent-sample t-test while Mann-Whitney U or Wilcoxon signedrank tests was used to compare abnormal distributions between two groups (responders and non-responders). A one-way analysis of variance (ANOVA) for repeated measurements has been performed for analysis of the hemodynamic values from T 0 to T 3 . The Tukey test was used for post hoc pairwise multiple comparisons analysis. Moreover, the predictive value of SVV, PPV, TVC and FC were evaluated by receiver operating characteristic (ROC) curve (95% con dence interval). The DeLong test was used to compare the statistically signi cant ROC curves [15]. Statistical analyses and graphics were conducted using GraphPad PRISM V7 (GraphPad Software Inc, San Diego, CA). We considered a P values less than 0.05 to be signi cant for all comparisons.

Results
100 patients scheduled for laparoscopy -assisted radical gastrectomy under general anesthesia were screened and 90 considered eligible in the enrollment period. Totally 9 patients were excluded with following reasons: 8 patients were not accordant with the inclusion criteria and 1 suddenly decided to give up the operation before arrival to the operating room). 81 patients were eventually included, 37 were responders and 44 patients were nonresponders, grouped based on the administration of FC. (Fig. 2)

Preoperative characteristics
As shown in Table 1, there was no statistically signi cant difference between the two groups in the baseline characteristics (gender, age, BMI, ASA, and preoperative complications), respiratory characteristics (peak airway pressure, driving pressure, oxygenation index, P pl and C rs ) and hemodynamic characteristics (MAP, central venous oxygen saturation and the parameters of arterial blood gas analysis ) (P > 0.05).

Effect of TVC and FC administration
The TVC signi cantly increased PPV (from 7.5-11.2%, P < 0.001) and SVV (from 8.6-13.4%, P < 0.001) in responders, but didn't affect PPV and SVV in non-responders (T 1 vs T 0 ). PPV and SVV in the responder's group were signi cantly higher than those in the non-responder's group at T 1 ( Table 2). Following FC administration, there were signi cant decrease in PPV (from 12.0-5.5%, P < 0.001) and SVV (from 14.3-6.5%, P < 0.001) only in responders. And FC also signi cantly increased CVP only in responders. However, the all recorded variables in non-responders, as shown in Table 2, haven't been changed by the administration of FC.

Prediction of Fluid Responsiveness
As shown in Table 3, the PPV 6 and SVV 6 did not correlate to uid responsiveness while ΔPPV 6 − 8 , ΔSVV 6    The box and whisker plots have been performed to compare these variables (ΔPPV 6 − 8 , ΔSVV 6 − 8, ΔPPV fc and ΔSVV fc ) among two groups (responders and non-responders) and there was a signi cant (P < 0.05) difference (Fig. 4). The ROC curves of above variables were shown in Fig. 3.

Discussion
The accuracy of predicting uid responsiveness in patients is the key to guiding perioperative uid management [16]. Therefore, it is necessary to identify the most accurate measures of uid responsiveness with the objective of minimizing incidents of uid overload and hypovolemia. The PPV could show predictive value with a Vt at least 8 mL/kg PBW as shown in the study of De Backer D et al [14]. Several studies have indicated that the SVV and PPV may signify a nonresponsive status even in responders during low Vt ventilation. The reason is that the Vt might be insu cient to produce a signi cant change in the intrathoracic pressure [17][18][19]. Meanwhile, previous studies show that ventilation using PCV-VG combined with low Vt (6-8 ml/kg PBW) has been effective as a lung protective ventilation strategy [10,11]. It is not clear whether PPV and SVV can accurately assess patients' uid responsiveness during mechanical ventilation using PCV-VG combined with low Vt (6-8 ml/kg PBW). In recent years, the assessment of uid responsiveness performed via dynamic evaluation of hemodynamic parameters in response to certain interventions, known as functional hemodynamic tests, such as mini uid challenge test (MFT), tidal volume challenge (TVC) and end-expiratory occlusion test (EEOT), have been considered reliable and effective methods in guiding perioperative uid management [20][21][22].
This study initially explored the reliability of functional hemodynamic tests in predicting uid responsiveness in patients ventilated using PCV-VG. The main nding showed that ΔPPV 6 − 8 and ΔSVV 6 − 8 are remarkable predictors of uid responsiveness in patients undergoing laparoscopy-assisted radical gastrectomy with cutoff values of both 1.5%. The change in PPV and SVV after a uid challenge (ΔPPV fc and ΔSVV fc ) also accurately predicts uid responsiveness with very high sensitivity and speci city. And yet, it requires a uid bolus to discriminate responders from non-responders seen in the change of cardiac index and may increase the risk of uid overload in the non-responder corhort. PVV and SVV at Vt 8 mL/kg PBW also identi es responders with the area under receiver-operating characteristic curve (AUC) (0.87 and 0.85, respectively), which are lower than those of ΔPPV 6 − 8 and ΔSVV 6 − 8 (Table 4). Therefore, TVC should be a good strategy to enhance the predictive value of PPV and SVV for the evaluation of uid responsiveness in patients undergoing protective ventilation with small tidal volume.
Meanwhile, many studies have shown that using PCV-VG can reduce lung injury caused by mechanical ventilation as well as reduce expiratory pressure and improve arterial oxygenation. This is assumed to be a bene cial effect of the decelerating ow rate on the airway and the decelerating waveform on intrapulmonary distribution by PCV-VG [23][24][25]. Therefore, we suspected that the decelerating ow rate may in uence the change of patients' intrathoracic pressure and the cardiopulmonary interactions weakly, thus affecting the accuracy of the dynamic indicators. The TVC (an increase of V t from 6 to 8 mL/kg PBW) can reduce the in uence of the decelerating ow rate on the change and conduction of intrathoracic pressure. As a result, ΔPPV 6 − 8 and ΔSVV 6 − 8 accurately predicted uid responsiveness depending on the increased uctuation of intrathoracic pressure, as well as greater dynamic compliance and cardiopulmonary interaction, during the implementation of TVC. This was con rmed in this study: In line with the ndings of previous studies conducted in patients ventilated with the VCV mode, this study showed that the uid responsiveness of patients predicted by the PPV 6 and SVV 6 under ventilation of PCV-VG mode were also not valuable. But the AUCs of PPV 6 , SVV 6 in this study is lower than those in previous studies.
Moreover, the AUCs of PPV 8 and SVV 8 are also lower than those in studies conducted in patients ventilated with the VCV mode. Nevertheless, it has been showed that the AUCs of ΔPPV 6 − 8 and ΔSVV 6 − 8 in this study were not signi cantly different with those in previous studies [26][27][28]. Simultaneously, the correlation between the ΔPPV 6 − 8 and ΔSVV 6 − 8 after the TVC application and ΔCI after FC administration (Fig. 5) suggests that the using PCV-VG with low tide volume in surgical subjects with normal respiratory compliance may not alter the interaction between volume status, the transmission of the intrathoracic pressure to the heart and the nal effect on PPV and SVV, ensuring the application of TVC suitable for the setting of PCV-VG mode. The results shown above could support our hypothesis and con rm that the absolute changes in PPV and SVV (ΔPPV 6 − 8 and ΔSVV 6 − 8 ) after TVC could also be a reliably functional hemodynamic test for predicting uid responsiveness when using PCV-VG combined with low V t (6-8 ml/kg PBW).
Our research showed that the TVC signi cantly increased PPV and SVV in responders but didn't affect PPV and SVV in non-responders, which is consistent with previous reports about TVC. The threshold of ΔPPV 6 − 8 and ΔSVV 6 − 8 were lower than those initially reported by Myatra et al [6]. This nding may be explained by two factors.
On the one hand, compared with VCV mode used in Myatra's study, the decelerating ow rate of PCV-VG may weaken the transmission of intrathoracic pressure to pleural and atrial pressure and in uence cardiopulmonary interaction and lead to a decrease of threshold of these variables. On the other hand, different from the patients with ASA physical status I-II and normal lung compliance we selected, the patients in Myatra's trail were critically ill patients with acute circulatory failure, 30% of whom were affected with a reduced chest wall compliance, which may enhance the transmission of applied airway pressure to the pericardium and the vena cava [28,29]. In other words, there were varying degrees of in uence of TVC application in elective surgical and critically ill patients on account of the distinction of cardiopulmonary interactions [30].
Some limitations of this study should be discussed. First, the time frame takes place after the induction of anesthesia and before the intervention of surgery, because the aim was to evaluate whether the TVC could enhance the predictive value of PPV and SVV in patients under the ventilation of PCV-VG. Therefore, it is impossible to predict the in uence of different surgical types and operations on the results, which still needs further study; Secondly, this study only explored the reliability of PPV and SVV in predicting uid responsiveness under ventilation of PCV-VG, but the reliability of other functional hemodynamic tests, such as EEOT and MFT, are not clear and also need further study. Finally, the other limitations in application of PPV and SVV in ventilation of PCV-VG, such as cardiac arrhythmias, the presence of spontaneous breathing and right ventricular dysfunction, could not be avoidable.

Conclusions
A transient increase in tidal volume, which we reference as the "tidal volume challenge (TVC)" could enhance the predictive value of PPV and SVV for the evaluation of uid responsiveness in patients under ventilation with PCV-VG. Written informed consent was obtained from the all participants or guardians before surgery.

Consent for publication
Not applicable.

Availability of data and materials
The datasets generated and / or analyzed during the current study will be available from the corresponding author on reasonable request.

Con ict of interest
The author declares that there is no con ict of interest that could be perceived as prejudicing the impartiality of the research reported.   The ow diagram of study.