DOI: https://doi.org/10.21203/rs.3.rs-1442276/v1
Background. The mean systemic filling pressure (Pms) play a central role for the understanding of the circulation. We studied whether the Pms indicates fluid responsiveness before and after induction of general anesthesia and whether the Pms response to bolus infusions of fluid indicates an anesthesia-induced increase of the "unstressed" blood volume.
Methods. An analog to Pms based on cardiac output, the mean arterial pressure, and central venous pressure (Pmsa) was calculated in 86 patients before induction of general anesthesia and before each of 3 successive bolus infusions of 3 ml/kg of colloid fluid. An increase in stroke volume of ≥ 10% from a bolus infusion indicated fluid responsiveness. Changes in blood volume were estimated from anthropometric data and the hemodilution.
Results. Pmsa was lower in fluid responders than in non-responders before induction (13.2 ± 2.2 vs. 14.7 ± 2.7 mmHg; mean ± SD, P< 0.01) and after induction of general anesthesia (11.4 ± 2.1 vs. 12.8 ± 2.1 mmHg; P< 0.006). Changes in Pmsa resulting from the infusions did not differ depending on the stroke volume response. Receiver operator characteristic curves showed an average area under the curve of 0.70. The decrease in Pmsa due to the anesthesia-induced increase of the vascular compliance was fully compensated by the colloid fluid. A linear correlation between Pmsa and the volume changes suggested that the anesthesia increased the "unstressed" blood volume by 1.2 L.
Conclusions. Pmsa was lower in fluid responders than in non-responders. General anesthesia increased the need for blood volume by 1.2 L.
The mean systemic filling pressure (Pms) is the pressure that develops in the systemic circulation if the heart suddenly stops [1]. The importance of Pms for the vascular status was first studied by the British physiologist Arthur Guyton. His view was that the heart fills passively. Therefore, cardiac output (CO) is determined by the venous return (VR), which is, in turn, driven by the difference between Pms and the central venous pressure (CVP) divided by the resistance to venous return (RVR). The theories surrounding the role of Pms as a key determinant of the circulation are sometimes called “Guyton´s hemodynamics” and offer complementary views on how to interpret hemodynamic data [2, 3].
A key problem is that Pms is difficult to measure, which necessitated the development of predictive algorithms. The best-known analog, Pmsa, is based on CVP, mean arterial pressure (MAP), and CO [4, 5]. This analog is implemented in a commercially available monitor, Navigator (Applied Physiology, Pty Ltd., Sydney, Australia). Cecconi et al. connected a Navigator module to a pulse contour hemodynamic monitor and recorded Guyton´s variables in postsurgical patients. Their reported hemodynamic changes agreed with Guyton´s views [6]. Later evaluations showed that Pmsa correlates well with more invasive laboratory methods of measuring Pms [7–11].
The purpose of the present report was to evaluate if the Pmsa predicts whether a patient is fluid responsive. The assessment of fluid responsiveness is the key methodology used for clinical evaluation of the need for fluid administration during surgery and intensive care. A secondary purpose was to estimate the increase in "unstressed blood volume" that occurs when general anesthesia is induced. To our knowledge, this is a novel use of Pmsa.
The hypotheses were that Pmsa predicts fluid responsiveness and that data on Pmsa can provide information about the "unstressed" blood volume.
This is a retrospective analysis of a prospective study that included patients with suspected or established gastric, colonic or rectal cancer who were recruited to participate in an open-labeled clinical trial [12, 13]. They underwent laparoscopic or open gastrointestinal surgery under combined intravenous and inhalational general anesthesia. Exclusion criteria were liver or renal dysfunction (liver enzymes > 50% or serum creatinine > 50% of normal), coagulation disturbances, obstructive pulmonary disease, atrial fibrillation, and mental disorders.
The protocol was approved by the Ethics Committee of the First Affiliated Hospital, College of Medicine, Zhejiang University (Hangzhou, PR of China; No. 2011150, Official in charge: Zhangfei Shou) and the registered at the Chinese Clinical Trial Registry (http://www.chictr.org/en; No. ChiCTR-TNRC-14004479). Written informed consent was obtained from each study subject. Reporting adhered to the CONSORT checklist.
The patients fasted overnight and no premedication was given. Anesthesia was induced and tracheal intubation performed by using propofol, fentanyl, and cisatracurium. Mechanical ventilation was used with a tidal volume set to 8 ml/kg, 12 breaths/min, and a PEEP of 3 cm H2O. The anesthesia was maintained by 1–2% of sevoflurane and infusion and remifentanil. No adrenergic drugs were administered.
No fluid was induced during the induction of general anesthesia. Beginning 10 min after the tracheal intubation, three bolus infusions of 6% hydroxyethyl starch 130/0.4 (Voluven®; Fresenius Kabi Deutschland GmbH, Bad Homburg, Germany) were given in the volume at 3 ml/kg over 7 min via an infusion pump (IEC 601–1; Abbott Laboratories, Chicago, IL). The hemodynamic response was recorded 5 min after the end of each bolus infusion. The flat recumbent body position was maintained, and surgery was not initiated until all three optimizations had been completed [12].
When the patient entered the operating theater, catheterization of the left radial artery and right interval jugular vein was performed under local anesthesia and sedation by midazolam. The arterial line was connected to a FloTrac™ sensor, from which data were sent for analysis to a Vigileo monitor (Software version 3.6; Edwards Lifesciences, Irvine, CA). The arterial waveform pulse contour was used to calculate the stroke volume (SV). Monitoring also included central venous pressure (CVP), pulse oximetry, electrocardiography, and heart rate (Datex-Ohmeda, Hoevelaken, the Netherlands).
The CVP was calibrated prior to induction of anesthesia. The zero point corresponded to the level of 4th rib in the anterior axillary line. The effect of a few extreme outliers was reduced by setting changes in CVP > 4 mmHg in response to a single bolus infusion at 4 mmHg.
Data on central hemodynamics were collected before and after induction of anesthesia, just before the first bolus infusion was initiated, and again 5 min after each of the bolus infusions ended.
The target in flow-guided optimization with fluid loading is to reach the top of the Frank-Starling curve. Therefore, the patient is a responder if a bolus infusion raises SV by ≥ 10% and non-responder if the increase is < 10% [14, 15]. As flow-guided optimization implies a titration process, a bolus is indicated if given after an infusion in which the patient was fluid responsive, but the subsequent bolus is warranted only if the SV increased by ≥ 10%.
An analog to the mean circulatory filling pressure (Pmsa) has been derived from measurements of CVP, MAP, and CO, assuming a constant veno-arterial compliance of 24:1 [4–6]:
Pmsa = a CVP + b MAP + c CO
where a = 0.96 and b = 0.04 (a + b = 1) while c is a resistance derived from anthropometric data [6];
c = 0.0038 (904.17 + 0.193 age) / [4.5 (0.99 age−15 0.007184 (height0.725) weight0.425]
Pressure gradient for venous return (dVR) is obtained as: dVR = Pmsa – CVP
The global pumping efficiency (Eh) is calculated as: Eh = (Pmsa – CVP) / Pmsa
The resistance to venous return (RVR) was obtained as: RVR = dVR / CO
The blood volume changes in response to the bolus infusions were calculated by multiplying the change in the blood hemoglobin concentration with the baseline blood volume, which was estimated based on the height and weight of each volunteer [16].
The data are presented as mean (SD) and differences between groups evaluated by one-way analysis of variance (ANOVA). P < 0.05 was considered statistically significant.
Receiver operator characteristic (ROC) curves were created with IBM SPSS Statistics Version 22. The ROC curves are probability curves in which sensitivity (true positive fraction) is plotted versus 1 – specificity (false positive fraction). The calculated area under the curve (AUC) for this relationship reflects how well the ranges of fluid intake can be separated. The given prediction is statistically significant if the 95% confidence interval does not include 0.5.
The cohort consisted of 86 patients (65% male). Data were missing from 7 patients, so the final analysis consisted of 79 subjects. They subjects were 56 ± 13 years old, had a height of 184 ± 8 cm, and body weight of 60 ± 8 kg. All patients received three bolus infusions after general anesthesia had been induced. The hemodynamic data are summarized in Table 1.
|
Before anesthesia |
Before 1st bolus |
Before 2nd bolus |
Before 3rd bolus |
After 3rd bolus |
|
|---|---|---|---|---|---|
|
Stroke volume (ml) |
82 ± 25 |
53 ± 16 |
60 ± 15 |
65 ± 16 |
67 ± 16) |
|
MAP (mmHg) |
104 ± 13 |
76 ± 10 |
75 ± 10 |
74 ± 11 |
75 ± 10 |
|
CVP (mmHg) |
5.0 ± 3.0 |
6.1 ± 3.3 |
6.7 ± 3.2 |
7.4 ± 3.1 |
8.3 ± 3.2 |
|
Pmsa (mmHg) |
13.8 ± 2.5 |
11.9 ± 2.2 |
12.4 ± 2.2 |
13.1 ± 2.3 |
14.0 ± 2.4 |
|
dVR (mmHg) |
9.0 ± 1.1 |
4.8 ± 1.8 |
5.3 ± 1.2 |
5.2 ± 1.3 |
5.1 ± 1.2 |
|
Eh (no unit) |
0.67 ± 0.11 |
0.42 ± 0.18 |
0.44 ± 0.12 |
0.41 ± 0.13 |
0.37 ± 0.11 |
|
RVR (mmHg min/L) |
1.5 ± 0.3 |
1.3 ± 0.4 |
1.4 ± 0.4 |
1.4 ± 0.4 |
1.3 ± 0.3 |
|
VR (L/min) |
6.3 ± 1.7 |
3.8 ± 1.1 |
3.9 ± 1.0 |
4.0 ± 1.2 |
3.9 ± 1.1 |
|
”Warranted” bolus |
63% |
44% |
22% |
Fluid responsiveness was evident in 63%, 44%, and 22% of the patients before each bolus infusion. The ROC curves showed that Pmsa could separate responders from non-responders with an AUC of approximately 65–70% before administration of any of the bolus infusions. Fluid responsiveness could be indicated even before anesthesia induction (Fig. 1).
Pmsa measured before the induction of anesthesia indicated how many of the bolus infusions that would later become "warranted" (P < 0.03; Fig. 2).
Pmsa differed significantly between subjects who would be non-responders and those who would be responders during the subsequent bolus infusion. This was a consistent finding (Fig. 3A). Induction of anesthesia was followed by a marked decrease in both dVR and Eh, but further changes and differences between non-responders and responders were negligible (Fig. 3B, C). RVR tended to be higher in the responders, but differences were small (Fig. 3D). Stroke volume showed the same pattern as Pmsa, but the differences between responders and non-responders were smaller for stroke volume (Fig. 3E).
To understand Fig. 3, note that the patients were continuously redefined as non-responders and responders and that each patient could switch between these groups at different points in time.
"Unstressed volume"
Figure 4 illustrates how the "unstressed blood volume" was increased by general anesthesia. The onset of anesthesia decreased Pmsa by 3.3 mmHg, as indicated by an arrow. The blood volume at baseline amounted to 4.3 ± 0.8 L, and the volume expansion from the bolus infusions is plotted versus Pmsa to obtain a vascular compliance curve. The increase in the "unstressed blood volume" amounted to 1.2 L, which is indicated by the horizontal shift from the baseline Pmsa (13.8 mmHg) to this curve.
Pmsa predicted fluid responsiveness before a fluid bolus was infused during general anesthesia. Pmsa also indicated the fluid responsiveness prior to anesthesia induction and even how many fluid boluses would be needed until SV no longer increased by ≥ 10%. However, the overall discriminating capacity of Pmsa to predict fluid responsiveness was not impressive. The ROC curves yielded confidence intervals that were statistically significant but only with modest margins.
We used the fluid-induced responses in Pmsa to estimate how much the anesthesia-induced reduction in the vascular compliance increased the "unstressed" blood volume, which is the fraction of the intravascular volume that does not generate pressure [3]. Fig. 4 shows that 1.2 L of blood would be needed to restore Pmsa to its pre-anesthesia level.
The research works by Arthur Guyton from the 1950s link circulatory volume with pressure and flow. The central concept is the mean circulatory filling pressure (Pmcf), which is the pressure that in develops in vascular system if the blood flow is quickly stopped. A closely related variable is the mean systemic filling pressure (Pms) which denotes the pressure when equilibrated throughout the systemic circulation [2]. The Pms and Pmcf values are usually similar and are often used interchangeably.
The driving force for venous return (dVR) is the difference between Pms and the right atrial pressure, which is measured clinically as the CVP. Thus, a high CVP operates as a resistor to the venous return, which governs CO. The flow gradient is stronger if Pms is high, which is expressed by the parameter denoted Eh, because the resistance to flow by the CVP then becomes less important. One may say that Eh is a measure of how effectively a volume change increases the CO.
The likelihood of fluid responsiveness was higher when Pmsa was low. This is logical, as Pmsa reflects the "stressed" blood volume and the vascular compliance. The fluid-induced increases in Pmsa did not differ significantly between responders and non-responders, and this was also expected because the same fluid volume was given to all patients.
The dVR and Eh decreased by 30–40% after the induction of general anesthesia, which was reflected in a drop in stroke volume by 35%. By contrast, Pmsa only decreased by 25%, as shown in Fig. 4. This difference can be explained by the increase in CVP, which is mostly likely due to the positive-pressure that was initiated as soon as patients were anesthetized. This suggests that 2/3 of the reduction of the SV could be accounted for by an anesthesia-induced increase in the vascular compliance. The resistance to venous return (RVR) is not expected to change during hyper- or hypovolemia [8], and only small changes were found in the present study.
The "unstressed" blood volume is usually obtained from the intercept of the y-axis (volume) at zero pressure in a vascular compliance plot [3]. Fig. 4 shows this type of plot, but it is based on only the changes within the narrow interval of the present measurements. Here, the horizontal shift between the baseline Pmsa and the compliance curve indicated the increase in the "unstressed" blood volume due to general anesthesia. This is the volume that the anesthetist aims to compensate using intravenous fluid.
The particularly pronounced blood volume response to the first bolus infusion is probably due to the capillary refill that always occurs in response to the decrease in arterial pressure accompanying anesthesia induction, even in the absence of intravenous fluid administration [17]. Capillary refill is also the reason why the first post-induction Pmsa could not be used in Fig. 4, as the blood volume change was not zero and no matching hemoglobin value was available.
Overall, the bolus infusions expanded the blood volume by more than the infused amount. This is reasonable, as the colloid osmotic pressure of the fluid is 33% higher than normal blood plasma [18]. However, the intravenous retention of the infused fluid, being higher for colloids than for crystalloids, is unlikely to matter much for the present calculations.
Crystalloid fluid might even offer an alternative way to estimate the anesthesia-induced increase of the "unstressed" volume. Kinetic analysis of hemodilution curves in women scheduled for abdominal hysterectomy showed that capillary leakage of fluid was arrested when 1.24 L (16.6 ml/kg) of Ringer´s had been infused, which is similar to the value found here [19]. This finding suggests that a low Pmsa counteracts the capillary leakage of fluid when the blood volume is expanded by crystalloid fluid.
The central idea of Guyton´s hemodynamics is that CO is determined by the venous return, whereas the heart plays a permissive role [3]. Despite critical views, this concept has received widespread attention among physiologists, anesthetists, and intensivists alike [20].
Basic studies have been performed in pigs, where Pms has been derived by ventilatory maneuvers [21, 22] and, recently, with extracorporeal membrane oxygenation [11].
Attempts to use Pms in the clinic have been made over the past decade. Three methods are used. One is to calculate Pms when VR is suppressed by stepwise deep inspirations. The second is to arrest the circulatory flow in one arm by inflating a blood pressure cuff and then to obtain Pms when the arterial and venous pressures have become equal. The third method is to calculate the Pms analog called Pmsa, which was the approach used in the present work.
Comparisons between these methods in cardiac surgery have shown, in one study, acceptable agreement between Pms values, but good agreement between changes in effective blood volume [7]. Meijs et al. compared inspiratory holds with Pmsa in cardiac surgery and found the methods to be interchangeable [10].
Cecconi et al. measured Guyton´s hemodynamic parameters in 39 postoperative patients who received different vasoactive therapies and respiratory support [6]. The Pmsa showed great variability and did not increase in response to a fluid bolus consisting of either crystalloid or colloid fluids.
A review by Cooke et al. supports our finding that Pmsa is lower in fluid responders than in non-responders [9]. However, fluid challenges and passive leg raising increased Pmsa more in the responders, and by a greater incremental change than we found. These differences may be due the anesthetized state of our patients.
CO was measured by the arterial waveform pulse contour analysis implemented in the FloTrac/Vigileo system. This hemodynamic monitor can be used in conscious patients, but it is uncalibrated and may then have a higher coefficient of variation than is observed with calibrated monitors [23].
The data were collected in the clinical setting, and the occasionally high SV apparent at baseline may be due to preoperative stress.
The strengths of the study include the uniform anesthetization of the patients and their freedom from acute disease. All patients also received the same fluid treatment. Sampling was carefully timed by a single set of investigators. No adrenergic drugs were used, as they may affect vascular tone and Pmsa [24].
The original patient series included 25 additional patients who received volume loading with Ringer´s solution in addition to the 86 who received colloid. Those who received Ringer were not included, as 3 ml/kg of crystalloid could not adequately challenge fluid responsiveness [12]. Only 20% of these patients showed fluid responsiveness during the first bolus infusion, which is 1/3 of the fraction of patients who received the colloid bolus.
A mean systemic filling pressure analog (Pmsa) indicated fluid responsiveness in patients who were given general anesthesia followed by three successive bolus infusions of colloid fluid. A comparison between the changes in Pmsa and the estimated blood volume changes suggested that general anesthesia increased the "unstressed blood volume" by as much as 1.2 L.
AUC: area under the curve; BIS; bispectral index; CO: cardiac output; CVP: central venous pressure; dVR: pressure gradient for venous return; Eh: global pumping efficiency (Eh); PEEP: positive end-expiratory pressure; Pms: mean systemic filling pressure; Pmsa: mean systemic filling pressure analog; Pmcf : mean circulatory filling pressure; ROC: Receiver operator characteristic; RVR: resistance to venous return; SV: stroke volume; VR: venous return.
Acknowledgements. Dr. Xiaojiang Ying recruited the patients. Operation theater nurse Guofang Meng assisted during the experiments.
Author's contributions. YL and RGH planned the study. YL rote the applications and arranged the funding. RH and YL collected the data. RGH made the calculations and authored the manuscript.
Funding: This project was funded by Qianjiang Talents Project of the Technology Office in Zhejiang province (No. 2012R10033), PR of China, and by a grant from the Östergötland City Council (No. LiO-297751), Sweden.
Availability of data and materials. The data used for the statistics is appended as Additional File 1.xls. All original is given in Additional File 2.xls.
Ethics approval. The protocol was approved by the Ethics Committee of the First Affiliated Hospital, College of Medicine, Zhejiang University (Hangzhou, PR of China; No. 2011150, Official in charge: Zhangfei Shou) and the registered at the Chinese Clinical Trial Registry (http://www.chictr.org/en; No. ChiCTR-TNRC-14004479). The study was performed in accordance with the Declaration of Helsinki. Written informed consent was obtained from each study subject.
Consent for publication. Not applicable.
Competing interests. RGH has received a research grant from Grifols for studies of 20% albumin and is Member of Baxter´s IV Fluid Therapy management Advisory Board.
YL and RGH have no conflicts of interest to report.
ORCID ID Robert G. Hahn 0000-0002-1528-3803