Administration of IV crystalloid fluid is a cornerstone of perioperative care (1). Many patients still receive fluid according to rigid conventional programs, but several outcome studies have recently shown that individual optimization might be a better choice (2,3). Hospital stay can be shortened by goal-directed fluid therapy, which means that the cardiac output response to repeated small doses of a colloid fluid is tested in the individual patient. Intravascular fluid administration is considered optimal before surgery when no increase in cardiac output is obtained with additional fluid (4–6). Volume kinetics, a branch of pharmacokinetics, has more fully clarified when and how distribution and slow elimination of a crystalloid enhance the volume effect of infused fluid (7,8).
The hypothesis in the present study is that plasma dilution can serve as a link between infused fluid and hemodynamic response during surgery, just as plasma concentration of a drug is an intermediate between administered drug therapy and its effect (9). If this is the case, an IV infusion of a crystalloid would generate a predictable plasma dilution and, therefore, a predictable hemodynamic effect. To study this, we examined a bolus infusion of a crystalloid during the first 75 min of upper abdominal surgery. These procedures normally require invasive hemodynamic monitoring on clinical grounds and are associated with an initially minor and predictable bleeding. We continuously measured hemodynamic response by using a pulmonary artery wedge catheter and also took frequent blood samples to measure the plasma dilution. Remifentanil was used as the key anesthetic to prevent undue depression of the cardiovascular system.
Ten elective surgical patients, ASA physical status II (two ASA III), were recruited after being scheduled for major abdominal surgery under general anesthesia. The protocol had been approved by the appropriate IRB, and each patient gave his or her written consent to participate. There were 5 men and 5 women aged between 46 and 73 yr (mean, 58 yr), with a body weight of 55–85 kg (mean, 69 kg), included either at the Stockholm Söder Hospital in Stockholm, Sweden, or the University of Texas Medical Branch at Galveston, TX. Six patients suffered from pancreatic cancer, two from pancreatitis and pancreatic cyst, respectively, and the last two from unclear abdominal pain.
The patients were bowel prepared on the day before surgery. The patients were fasting from midnight and given 2.5 mg/mL of glucose solution, the amount corresponding to (weight in kg + 40) milliliters per hour.
Patients were given 2–5 mg of midazolam IV. Monitoring consisted of electrocardiogram, pulse oximetry, and noninvasive arterial blood pressure. Two 16-gauge IV catheters (Becton Dickinson, Sandy, UT) were placed in antecubital veins, one on each side. One 20-gauge arterial catheter, either Abbocath®-T (Abbot Ireland, Sligo, Ireland) or Arrow International (Arrow International, Reading, PA), was also placed in the left radial artery under local anesthesia and aseptic conditions. A 19-gauge thoracic epidural catheter (Arrow International) was also placed at level T8 and tested but not activated.
When the patient was still awake, one percutaneous 8.5F sheath introducer (Arrow International) was inserted into the right internal jugular vein. A Swan Ganz CCOmbo V CCO/SvO2/CEDV/VIP Thermodilution Catheter (Edwards Lifesciences LLC, Irvine, CA) was floated and connected to a Vigilance™ CEDV monitor (Edwards Lifesciences). A Foley catheter was placed in the bladder for measurement and sampling of the urinary excretion.
Anesthesia was provided by using propofol 10 mg/L, glycopyrrolate 0.15 mg, and remifentanil 0.50–0.75 μg·kg−1·min−1. The anesthesia was maintained using remifentanil 0.10–0.20 μg·kg−1·min−1 and sevoflurane 0.5 minimum alveolar anesthetic concentration to avoid awareness. After tracheal intubation, which was facilitated by rocuronium, there was a 15-min period of surgical preparation without study interventions. When the surgeon was ready to make the first skin incision, an infusion of 25 mL/kg of lactated Ringer’s solution was started and administered at a constant rate over 45 min. The study ended after a follow-up of 30 min, during which surgery proceeded but no infusions were given. The blood loss was very small, less than <50 mL, in all patients during this early period of the operation.
Arterial blood pressure was measured continuously. Cardiac output, central venous pressure, pulmonary artery wedge pressure, and central venous oxygen saturation were measured. The wedge pressure was obtained by manually inflating the balloon-tipped catheter every 5 min. Samples were drawn in duplicate and the mean values used in subsequent calculations. Blood was sampled from the radial artery catheter during the study to measure the hemoglobin (Hb) concentration and the red blood cell (RBC) count in the routine hospital laboratory with a variation coefficient of 2%. Samples were drawn in duplicate and the mean values used in subsequent calculations. The plasma sodium concentration, the hematocrit (Hct), and plasma osmolality were measured in single samples. The total amount of sampled blood withdrawn was approximately 250 mL. Urine was also collected for measurement of the sodium concentration and the osmolality.
The volume kinetic model fitted to the blood and urine data implied that fluid infused at a rate ki expands a central body fluid space v, which strives to return to its baseline volume V (to regain body fluid steady state) by three mechanisms: first, translocation to a much larger peripheral body fluid space at a rate proportional by a constant kt to the dilution of V, i.e., the rate of translocation at any time was kt·(v−V)/V; second, elimination of fluid by urinary excretion at a rate proportional by a constant kr to the dilution of V, i.e., the rate of elimination at any time was kr·(v−V)/V; and third, a baseline loss (kb) fixed at 0.5 mL/min to account for evaporation from the surgical wound (10). Moreover, each withdrawal of blood was followed by an IV injection of 2 mL of normal saline to rinse the catheter and to replace fluid lost by evaporation from the airways and from the skin.
The volume change in the central body fluid space (v) was expressed as:
Input data comprised Hb, RBC, and Hct at baseline (time 0) and at any later time (t). They were used to calculate the plasma dilution, which also represented the dilution of V:
A correction of the plasma dilution for “iatrogenic” dilution resulting from the blood sampling was made based on the baseline blood volume (8) estimated from a regression equation using the height and weight of the patients as input data (11).
The unknown variables V and kt in the model were estimated by fitting the mathematical solution of the differential equation (Equation 1) to the data on plasma dilution (Equation 2). For this purpose, a nonlinear least-squares regression routine, based on a modified Gauss-Newton method, was repeated until no variable changed by more than 0.1% in each iteration. The variable kr was calculated from the urinary excretion; the total urine volume collected during the studied period was then divided by the area under the curve for the entire dilution-time profile. The amount of fluid residing in V was obtained as the product of the plasma dilution and the estimated value of V.
Results were presented as the mean and standard deviation or, if there was a skewed distribution, as the median and interquartile range. Groups of patients who had an increase in cardiac output in response to IV fluid (responders) were compared with those showing a decrease in cardiac output (nonresponders) by one-way or repeated-measures analysis of variance or Mann-Whitney test. P < 0.05 was statistically significant.
Infusion of crystalloid fluid increased cardiac output in four patients (responders) and decreased cardiac output in the other six patients (nonresponders). The mean body weight was virtually identical in these groups, being 68.3 and 69.5 kg, respectively. There was an overall increase in central venous and pulmonary artery wedge pressures during the fluid treatment (P < 0.001), although these pressures were higher in the responders (P < 0.05). The systemic vascular resistance decreased in the responders but increased in the nonresponders (Fig. 1). Central venous oxygen saturation remained stable at 75%–80% in both groups (data not shown).
Baseline values were Hb 11.43 g/dL (0.94), RBC 3.38 × 106 mm−3 (0.48), and Hct 35.33% (2.79). The plasma dilution increased gradually during the infusion and slowly returned to baseline after it ended (Fig. 2, top). However, the plasma dilution increased more among the responders, the difference being significant from the second half of the infusion and onward (P < 0.05; Fig. 2, bottom; two-way repeated-measures analysis of variance).
There was often a lateral shift of the linear correlation between hemodynamic change and plasma dilution during the infusion as compared with after the infusion. For example, cardiac output was relatively lower after the infusion ended among the responders, whereas it was higher for nonresponders (Fig. 3, top). The central venous pressures increased the same way in both groups, but the wedge pressure was lower for any specific degree of plasma dilution after the infusion in the group of responders (Fig. 3, middle and bottom).
Table 1 shows the results of the kinetic analyses and subsequent calculation of where the infused fluid resided at the end of the infusion and at the end of the study, respectively. The plasma and urinary electrolytes showed little change, and the patients concentrated the urine well (Fig. 4).
Our hypothesis received only partial support from the results. Fluid therapy resulted in a plasma dilution that was fairly predictable and could be quantified by a kinetic method, but the pharmacodynamic response was more or less inconsistent. The changes in central venous pressure and pulmonary artery wedge pressure did correlate well and consistently with the plasma dilution. However, fluid administration normally increases cardiac output and reduces the peripheral vascular resistance, and this occurred in only half the patients. The differences in response were already apparent after 5–10 minutes of the fluid administration and became gradually more pronounced during the first half of the 45-minute infusion. In contrast to awake sheep (9), the differences in slopes for the relationships between plasma dilution and cardiac output, and also between plasma dilution and peripheral vascular resistance, make these hemodynamic variables less useful as pharmacodynamic end points of fluid therapy.
It is unclear why half of the patients were nonresponders. A conventional and widespread view is that they were already at the top of the Frank-Starling curve, and further filling would be detrimental. Those who responded with an increase in cardiac output (responders) should therefore have been more hypovolemic. This interpretation forms the theoretical basis for goal-directed fluid therapy (5,6) and receives support from the seemingly lower cardiac output at baseline in the group of responders. This, however, was not statistically significant because of the small number of patients. However, the hypovolemia hypothesis can be questioned based on the fact that venous pressures were higher in the responders—a difference that remained throughout the experiment. Moreover, their urinary excretion tended to be larger.
All patients underwent enteric lavage before the operation, which normally causes slight dehydration (12,13). However, most of the fluid deficit should have been replaced by the maintenance administration of glucose solution during the night before surgery. Explanations other than differences in preoperative fluid status for the different cardiac output responses to intravascular fluid administration include variations in circulating catecholamines because of surgical stress. A further view is that the present findings might be caused by genetically determined variations in the cardiovascular adaptation to plasma volume expansion. These possibilities should be assessed in a larger study.
The hemodynamic response to intravascular fluid administration apparently accounts for much of the variability of the plasma dilution, which is an index of plasma volume expansion. The dilution differed significantly between responders and nonresponders but only from the second half of the infusion onward. The plasma dilution was similar during the first half of the infusion, which is the period of time when the hemodynamic changes were most variable. Interestingly, the process during which infused fluid fills up the increased vascular space that accompanies a reduced vascular tone also requires some time to be completed during the induction of general and regional anesthesia (14,15). As seen in the present and previous studies, a small spontaneous plasma dilution of 3%–4% develops during the induction of anesthesia if no IV fluid at all is administered (15,16). The volume of the IV anesthetics used for induction in the present study amounted to approximately 25 mL, whereas the accompanying plasma volume expansion was approximately 125 mL.
Plasma dilution served as the input variable for volume kinetic analysis. The model used here is a simplified form of the two-volume model (7). This is justified when the follow-up time is considerably shorter than the “normal” three hours required to also estimate the size of the peripheral body fluid space into which infused fluid is distributed (8). Overall, the size of the central body fluid space V was 3.25 L, which is the same as during laparoscopic cholecystectomy (7). However, the size of V tended to be smaller in the nonresponders, which is consistent with a difference in hemodynamic status between the groups (13). The intercompartmental clearance constant was 30% higher, and the elimination rate constant was three times higher, during open abdominal surgery as compared with laparoscopic cholecystectomy. In consequence, computer simulations suggested that the plasma dilution from crystalloid fluid infused during open abdominal surgery would be 30% less than during laparoscopy. However, the true elimination is not as rapid as might be suggested by the kinetics. The urinary flow rate during open surgery was 1.4 mL/min, whereas the corresponding rate for laparoscopy was 0.9 mL/min (7). The strong inhibition of urinary excretion during these operations, as well as during thyroid surgery (8), is evidenced by the elimination rate constant, which was only 15%–25% of that obtained in volunteers (17).
The similar sodium concentrations of the infused fluid and the excreted urine, as well as the unchanged serum sodium level, indicate that no marked fluid shifts between cells and noncells occurred. We have described that intravascular fluid administration with acetated Ringer’s solution in volunteers creates difficulties in excreting urine of the same sodium concentration as the infused fluid. This results in a net flow of fluid from the cells to the extracellular fluid space, which is explained by difficulties in concentrating rapidly excreted urine to approximately 500 mOsm/kg. Such elimination seems to be required to excrete enough sodium to prevent fluid redistribution (18). However, probably because the urinary excretion was so small, our patients undergoing open abdominal surgery had no problem with maintaining a high sodium excretion and adequately concentrating their urine.
The amount of fluid we infused, approximately 1.7 L, is probably quite normal practice in many hospitals for an induction of upper abdominal surgery and approximately 1–1.5 hours thereafter. Therefore, the hemodynamic pattern consistent with that of our nonresponders described here is probably quite common. It is uncertain whether fluid therapy offered any benefit at all to these patients. Nonresponders had a poor plasma volume expansion and were prone to accumulate the IV fluid in peripheral tissues.
In conclusion, half of the patients undergoing open abdominal surgery responded to crystalloid fluid with a decrease in cardiac output. Pulmonary artery wedge pressure and central venous pressure responded more consistently to different degrees of plasma dilution, which can be simulated for various fluid regimens using volume kinetics.
The authors are indebted to Roger Seeton, RN, UTMB, for blood sampling and analysis during the study.
1. Holte K, Sharrock NE, Kehlet H. Pathophysiology and clinical implications of perioperative fluid excess. Br J Anaesth 2002;89:622–32.
2. Nisanevich V, Felsenstein I, Almogy G, et al. Effect of intraoperative fluid management on outcome after intraabdominal surgery. Anesthesiology 2005;103:25–32.
3. Holte K, Klarskov B, Christensen DS, et al. Liberal versus restrictive fluid administration to improve recovery after laparoscopic cholecystectomy: a randomized, double-blind study. Ann Surg 2004;240:892–9.
4. Mythen MG, Webb AR. Intra-operative gut mucosal hypoperfusion is associated with increased post-operative complications and cost. Intensive Care Med 1994;20:99–104.
5. Mythen MG, Webb AR. Perioperative plasma volume expansion reduces the incidence of gut mucosal hypoperfusion during cardiac surgery. Arch Surg 1995;130:423–9.
6. Gan TJ, Soppitt A, Maroof M, et al. Goal-directed intraoperative fluid administration reduces length of hospital stay after major surgery. Anesthesiology 2002;97:820–6.
7. Olsson J, Svensen CH, Hahn RG. The volume kinetics of acetated Ringer’s solution during laparoscopic cholecystectomy. Anesth Analg 2004;99:1854–60.
8. Ewaldsson CA, Hahn RG. Kinetics and extravascular retention of acetated Ringer’s solution during isoflurane and propofol anesthesia for thyroid surgery. Anesthesiology 2005;103:460–9.
9. Ewaldsson CA, Vane LA, Kramer GC, Hahn RG. Adrenergic drugs alter both the fluid kinetics and the haemodynamic responses to volume expansion in sheep. J Surg Res. 2006;131:7–14.
10. Lamke LO, Nilsson GE, Reithner HL. Water loss by evaporation from the abdominal cavity during surgery. Acta Chir Scand 1977;143:279–84.
11. Nadler SB, Hidalgo JU, Bloch T. Prediction of blood volume in normal human adults. Surgery 1962;51:224–32.
12. Ackland GL, Singh-Ranger D, Fox S, et al. Assessment of preoperative fluid depletion using bioimpedance analysis. Br J Anaesth 2004;92:134–6.
13. Ewaldsson CA, Hahn RG. Volume kinetics of Ringer’s solution during induction of spinal and general anaesthesia. Br J Anaesth 2001;87:406–14.
14. Drobin D, Hahn RG. Time course of increased haemodilution in hypotension induced by extradural anaesthesia. Br J Anaesth 1996;77:223–6.
15. Hahn RG. Haemoglobin dilution from epidural-induced hypotension with and without fluid loading. Acta Anaesthesiol Scand 1992;36:241–4.
16. Holte K, Foss NB, Svensen C, et al. Epidural anesthesia, hypotension, and changes in intravascular volume. Anesthesiology 2004;100:281–6.
17. Hahn RG, Drobin D. Urinary excretion as an input variable in volume kinetic analysis of Ringer’s solution. Br J Anaesth 1998;80:183–8.
© 2006 International Anesthesia Research Society
18. Hahn RG, Drobin D. Rapid water and slow sodium excretion of acetated Ringer’s solution dehydrates cells. Anesth Analg 2003;97:1590–4.