Karolinska Institute, Department of Anaesthesia, Söder Hospital, S-118 83 Stockholm, Sweden*Corresponding author
Accepted for publication: March 26, 2001
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Br J Anaesth 2001; 87: 40614
Keywords: anaesthesia, subarachnoid; anaesthesia, general; fluids, i.v., Ringers solution; pharmacokinetics; kinetics
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In the present study, we use volume kinetic analysis to compare the distribution and elimination of Ringers acetate solution during the induction of spinal and general anaesthesia. The purpose of comparing these two situations is to obtain evidence as to whether the approaches for treatment should be the same. Furthermore, we wanted the kinetic analysis to possibly indicate how fluid can be used to prevent the hypotension.
Volume kinetics is based on the assumption that fluid given by i.v. infusion expands either one (V) or two (V1 and V2) fluid spaces in the body. The elimination of fluid from the system and the exchange of infused fluid between V1 and V2 are governed by the constants kr and kt, respectively.511 Repeated measurements of the blood haemoglobin (B-Hb) concentration and urinary excretion were used on a computer-model to find estimates of the parameters in the volume kinetic models.
![]() |
Patients and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
Enteric lavage, consisting of 4 litres of macrogol (Laxabon, Tika, Lund, Sweden) was given on the day before surgery to six patients scheduled for colorectal surgery. After fasting overnight, all patients underwent surgery, which started between 08:00 and 12:00. They received 57.5 mg of morphine by i.m. injection as premedication approximately 1 h before entering the operating theatre. With the patient in the supine position, a cannula was placed in the cubital vein of each arm for the respective purposes of sampling blood and infusing fluid. Fifteen minutes later, an i.v. fluid load of 20 ml kg1 of Ringer acetate solution (Pharmacia, Uppsala, Sweden) was given at a constant rate over 60 min (0.33 ml kg1 min1) via an infusion pump (Flo-Gard 6201, Baxter Healthcare, Deerfield, IL, USA). The Ringer solution had the following ionic content: Na 130, K 4, Ca 2, Mg 1, Cl 110 and acetate 30 mmol litre1.
Spinal (n=10) or general (n=10) anaesthesia was induced 20 min after starting the i.v. infusion. When spinal anaesthesia was induced, the patient was turned to a lateral position, and 23 ml of isobar bupivacaine 0.5% (Marcaine® spinal, AstraZeneca, Södertälje, Sweden) was injected. Immediately after injection, the patient was turned from the lateral back to the supine position. General anaesthesia was induced with thiopental 5 mg kg1 and maintained with fentanyl 0.15 mg (mean dose) and 13% sevoflurane in oxygen and ambient air. Endotracheal intubation was facilitated by i.v. injection of rocuronium 0.5 mg kg1.
Monitoring included pulse oximetry and electrocardiography. Non-invasive arterial blood pressure was measured every 3 min in the arm not used for infusion by an automatic device (Datex AS3, Datex, Helsinki, Finland). Ephedrine 510 mg was given as an i.v. bolus if the systolic arterial pressure dropped below 60% of baseline. Surgery was not started until the study period was completed.
Venous blood (3.0 ml) was collected every 3 min during the study period of 60 min. A small discard sample was drawn before each blood collection to preclude any admixture of blood from the previous sampling. The cannula was rinsed with 2.0 ml of saline after each sample collection to prevent clotting. The B-Hb concentration was measured by a Technicon H2 device (Bayer, Tarrytown, NY, USA) with a coefficient of variation of 1%. The first sample was drawn in duplicate and the mean value was used in the calculations. The excreted urine was collected via an indwelling catheter, which had been inserted into the bladder under topical anaesthesia before the study started.
Additional patients
After the main study had been completed, five additional patients between 55 and 85 (mean 71) yr of age with a body weight of 50103 kg (mean 77 kg) were studied to examine whether a specific fluid regimen suggested by the volume kinetic analysis of the first 20 patients is able to prevent arterial hypotension during the onset of spinal anaesthesia. The male/female ratio was 3/2 and the operations were similar as in the main study. In these additional patients, no preload was given but 350 ml of Ringers solution was infused during as little as 2 min just after injecting the bupivacaine. Thereafter, the fluid was given at the normal rate (0.33 ml kg1 min1) during 40 min. The same measurements were performed as in the main study.
Calculations
The distribution of the fluid given by i.v. infusion was analysed using volume-of-fluid-spaces kinetic models, which can be summarised as follows (Fig. 1, top).
|
In the two-volume model, the primary fluid space (v1) communicates with a remote fluid space (v2). The rate of volume equilibration between the expandable fluid spaces is proportional to the relative difference in deviation from the target values (V1 and V2) by a constant (kt) (Fig. 1, bottom). Here, the increased dilution associated with anaesthesia was assumed to represent a change in kt, for example, the model determined one kt before anaesthesia was induced and another for the time after the induction. The differential equations describing the volume changes in the expanded body fluid spaces as well as their matrix solutions are given in the Appendix.
The dilution of the plasma in the cubital vein was used to quantitate the water load as Ringers solution remains outside the erythrocytes. As the sampled plasma is a part of V, we obtain the following dilution at time t:
(v(t)V)/V=[baseline B-Hb/B-Hb(t) 1]/(1 baseline haematocrit)
A correction for the loss of erythrocytes during blood sampling was made based on a preoperative blood volume estimated from the patients height and body weight.5
The distribution of the infused i.v. fluid was modelled separately for each subject using Matlab version 5.2 (Math Works Inc., Notich, MA, USA), whereby a non-linear least-squares regression routine based on a modified Gauss-Newton method was repeated until no parameter changed by more than 0.001 (0.1%) in each iteration. The output of the kinetic analysis consisted in the best estimate and the standard error (SE) for V before and after induction (one-volume model) and for V1 and V2 as well as kt before and after induction (two-volume model). However, the results of the latter analysis were reported only if an F test indicated that it was statistically justified.6
The following assumptions were made to ensure sufficient stability of the models: (1) kr was calculated from the relationship between the integral of the dilution-time curve and the measured urinary excretion, both for the period of time up to when anaesthesia was induced and for the period thereafter.6 (2) kb was set to 0.8 ml min1 (700 ml per 24 h) which is a reasonable estimation of the basal fluid loss (insensible water loss and baseline diuresis) from V. Half of this figure was assumed to appear as urine. (3) In the two-volume model, the individual curves were analysed after using a fixed value of V2, assuming that the sum of V1 and V2 averages 12.5% of the patients body weight, which is an average figure found in previous studies.57
Data are presented as the mean and (SEM) except where noted. Differences within and between the groups were studied using the one-way and repeated-measures ANOVA. Correlations between parameters were evaluated by linear regression analysis. P< 0.05 was considered significant.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
The mean arterial pressure (MAP) tended to be higher before spinal anaesthesia (121 mm Hg) than before general anaesthesia (111 mm Hg) and the heart rate was lower (74 vs 83 beats min1), but these differences were not statistically significant (Fig. 3).
|
A search was made for factors correlating with the differences in kinetic parameters between individual patients. The decrease in MAP correlated with the size of V obtained after anaesthesia had been induced (one-volume model) and with the kt obtained after anaesthesia had been induced (two-volume model; Fig. 4).
|
Simulations and additional patients
Simulations outlined the volume increment representing, in the two-volume model, the increased vascular costume of V1 associated with anaesthesia (Fig. 5, left). This increased haemodilution required 20 min to develop fully, and averaged only 125150 ml. We hypothesized that such a volume increment would need to be filled up within a few minutes to prevent relative hypovolaemia in V1 during the onset of the anaesthesia, but this seemed impossible with the mode of infusion used in the study (Fig. 6). The goal could be reached only with a short rapid infusion, and the infusion rates required to fill V1 with various amounts of fluid within 2 min are shown in the right panel of Figure 5.
|
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
In most patients, the kinetics of the infused fluid was described according to a two-volume model, which means that anaesthesia alters the rate of volume equilibration, expressed by a parameter kt, between a central (V1) and a peripheral body fluid space (V2). The overall results indicate that the body handles infused fluid in a similar way during the induction of general and spinal anaesthesia. Both forms of anaesthesia were associated with a reduction kt that was also of similar magnitude. Such a decrease of kt quantitates an altered relationship in compliance between V1 and V2 with regard to volume expansion, which favours fluid accumulation in V1.
Another finding is that the size of V1 was very small. In volunteer experiments, V1 is usually close to the expected size of the plasma volume.57 Most or all of the plasma volume was probably a part of V1 during the initial 20 min of the present experiments too, but our partial derivative plots showed that the final estimate of V1 was strongly affected by the increased haemodilution occurring just after the onset of anaesthesia. A problem in the curve-fitting procedure was that V2 could not be estimated with confidence. Information about the size of V2 is best obtained from post-infusion data, which were absent this time. The solution to the problem we used, which consisted in assuming a fixed sum of V1 and V2 as suggested by other studies, introduces some uncertainty about the pre- and post-infusion estimates of kt, although their ratio is not changed.11 In the present study, however, the estimates of kt before the induction are similar to those obtained in volunteers.57 Furthermore, the final estimate of V1 is only marginally changed by setting different values for V1+V2.11
The combination of a small V1 and a reduced kt indicates that infused fluid circulates primarily in the central blood volume during the onset of anaesthesia. The resulting excessive dilution in V1 represents an enlarged vascular costume, which correlates with the magnitude of the arterial hypotension that develops during the onset of epidural12 13 and spinal14 anaesthesia. The maximum dilution during epidural anaesthesia is developed 1015 min after the drop in arterial pressure13 and the present study shows that the time lag is the same during induction of spinal anaesthesia. Hypotension developed somewhat later after the induction of general anaesthesia, which is probably a result of the stress associated with the intubation procedure.
Spinal anaesthesia allows relative hypovolaemia to develop in the torso by shifting blood to the legs in healthy volunteers15 and in patients with cardiac disease.16 The preferential distribution of infused fluid to the central plasma volume, together with the slower transport of fluid to a more remote body fluid space, is a meaningful adaptation, as infused fluid then restores cardiac preload more effectively. Under these conditions, the vasodilated legs may become part of V2. However, the infusion rates normally used for volume loading are simply too low to increase the volume of V1 by 125150 ml during the period of time between the injection of bupivacaine and the expected onset of hypotension 510 min later. Our data suggest that the strategy having the best chance to fill this increment, and which, therefore, might offer the best chance to prevent a drop in arterial pressure, would be to use a very high infusion rate just after the anaesthetic solution has been injected. Infusing fluid before the induction is less efficient as kt is higher at that time. In the additional series of five patients, we challenged this hypothesis about preventing hypotension by giving 2-min rapid fluid load just after the bupivacaine had been injected. Interestingly, no drop in MAP was observed, despite the fact that the analgesia was even more widespread in the additional patients than in the first 10 patients who received spinal anaesthesia. A follow-up study is planned to evaluate the clinical value of this fluid programme in a larger group of patients.
Another factor promoting central hypovolaemia when fluid is infused at a constant rate over 60 min fluid is that no recruitment of volume seems to occur from V2 to V1 during the onset of anaesthesia. Rapid transfer of fluid from muscle to the blood has been described when hypovolaemic stress is induced by a lower negative body pressure,17 but virtually no haemodilution could be observed when epidural anaesthesia was induced without being accompanied by fluid.12 Computer simulations based on the volume kinetic data derived here suggest that no such backward flow would occur even if more prolonged volume loading was performed before induction of anaesthesia (Fig. 6, lower). However, the linear relationship between kt after the induction and the relative change in MAP (Fig. 3, right) indicates that a transfer of fluid from V2 to V1 against a dilution gradient could occur when MAP falls to below 45% of baseline, as such a drop would theoretically result in a negative kt.
Four patients in each anaesthesia group had their dilution-time curves analysed according to a one-volume model. The two-volume model was discarded in these cases as a statistical test indicated that estimating three parameters instead of two did not result in a significantly reduced mean square error of the differences between measured and model-predicted data on dilution.5 Application of the two-volume model to these curves was possible but resulted in excessive SE for kt both before and after the induction. The dilution over time increased in a linear fashion as more fluid was infused, and the plots lack the exponential shape of the two-volume dilution-time curves.
No effect of the induction of anaesthesia became apparent in the one-volume plots. One possible effect would consist in a reduction of V after the induction, resulting in a more pronounced dilution, as was found before Caesarean section.11 A small V means that forces act to keep the infused fluid close to the bloodstream, similar to a low kt in the two-volume model, and is associated with anaesthesia-induced hypotension (Fig. 4, left).
Enteric lavage given on the day before the operation seems to have affected the present results. This was not expected when the study was started, but its importance became apparent when we sought an explanation for the difference between spinal and general anaesthesia with respect to the pre-induction size of V in the one-volume model. V was small enough to even correspond to the expected size of the plasma volume, which is a sign of hypovolaemia, and we, therefore, believe that the patients were dehydrated by the lavage. This view is supported by the slightly higher baseline heart rate and the more pronounced drop in arterial pressure in these patients. The fact that lavage was employed only in the general anaesthesia group is sufficient to explain the smaller size of V among these patients as compared with those that received spinal anaesthesia.
Another possible confounder in the study was the occasional use of a small dose of ephedrine to alleviate severe hypotension (systolic pressure <60% of baseline). The doses given are likely to have slightly and transiently reduced the differences in blood pressure response between spinal and general anaesthesia. Ephedrine acts by increasing the cardiac output and, to some degree, by causing vasoconstriction. As argued previously,11 this drug may have slightly reduced the vasodilatation and the rates of fluid exchange, but no indication of such effects could be obtained from the data. The effect of an i.v. dose of ephedrine is also fairly short. In contrast, the kinetic modelling was based on all data on haemodilution obtained during the 60 min of study.
An important finding is the weak diuretic response to volume loading. Only about 5% of the infused fluid volume of up to 2400 ml was excreted during the study period. The average half-life of Ringers acetate, as extrapolated graphically, was between 150 min (one-volume model) and 500 min (two-volume model) before anaesthesia. The induction further inhibited the diuretic response and prolonged the half-life. In comparison, the half-life was approximately 15 min in elderly volunteers.9 Preoperative stress and the overnight fast probably accounts for most of this difference. Interestingly, the diuretic response was stronger when the fluid therapy was started with a brisk 2-min infusion.
In summary, we found that induction of spinal and general anaesthesia results in similar changes in volume kinetic parameters and usually favoured an accumulation of infused fluid in a relatively small central body fluid space. The kinetic analysis suggests that the arterial pressure is better maintained by infusing fluid very rapidly just after the induction of anaesthesia than to give a preload.
![]() |
Acknowledgement |
---|
![]() |
Appendix |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
which is solved as a monoexponential solution. Before induction of anaesthesia, it is
and after (a) induction
where w(t) is the dilution (v(t) V)/V and V is the change in baseline (target) volume induced by the anaesthesia. kr is calculated from the measured urine excretion and has different values during and after the induction of the anaesthesia.
The following differential equations describes two-volume fluid space model:
As with the one-volume model, the equations have been adapted for the special situation during anaesthesia, and they calculate one kt for the period before and another kt for the period after the induction. The form used to present the solution of equations 4 and 5 is based on the matrix exponential eAt, which is implemented as a standard function in the mathematical program Matlab. Before anaesthesia, we have
and after (a) anaesthesia has been induced, it becomes
where the matrix A is
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
2 Nishimura N, Kajimoto Y, Kabe T, Sakamoto A. The effects of volume loading during epidural anaesthesia. Resuscitation 1985; 13: 319[ISI][Medline]
3 Coe AJ, Revenäs B. Is crystalloid preloading useful in spinal anaesthesia in the elderly? Anaesthesia 1990; 45: 2413[ISI][Medline]
4 Jackson R, Reid JA, Thorburn J. Volume preloading is not essential to prevent spinal-induced hypotension at Caesarean section. Br J Anaesth 1995; 75: 2625
5 Drobin D, Hahn RG. Volume kinetics of Ringers solution in hypovolemic volunteers. Anesthesiology 1999; 90: 8191[ISI][Medline]
6 Hahn RG, Drobin D. Urinary excretion as an input variable in volume kinetic analysis of Ringers solution. Br J Anaesth 1998; 80: 16[ISI][Medline]
7 Svensén C, Drobin D, Olsson, J, Hahn RG. Stability of the interstitial matrix after crystalloid fluid loading studied by volume kinetic analysis. Br J Anaesth 1999; 82: 496502
8 Hahn RG, Drobin D, Ståhle L. Volume kinetics of Ringers solution in female volunteers. Br J Anaesth 1997; 78: 1448
9 Svensén C, Ponzer S, Hahn RG. Volume kinetics of Ringer solution after surgery for hip fracture. Can J Anaesth 1999; 46: 13341[Abstract]
10 Svensén C, Hjelmqvist H, Hahn RG. Volume kinetics of Ringer solution during endotoxinemia in conscious rabbits. J Endotox Res 1997; 4: 42530[ISI]
11 Hahn RG, Resby M. Volume kinetics of Ringers solution and dextran 3% during induction of spinal anaesthesia for Caesarean section. Can J Anaesth 1998; 45: 44351
12 Hahn RG. Haemoglobin dilution from epidural-induced hypotension with and without fluid loading. Acta Anaesthesiol Scand 1992; 36: 2414[ISI][Medline]
13 Drobin D, Hahn RG. Time-course of increased haemoglobin dilution in hypotension induced by epidural anaesthesia. Br J Anaesth 1996; 77: 2236
14 Hahn RG. Increased haemodilution in hypotension induced by epidural anaesthesia. Acta Anaesthesiol Scand 1993; 37: 35760[ISI][Medline]
15 Arndt JO, Höck A, Stanton-Hicks M, Stühmeier K-D. Peridural anesthesia and the distribution of blood in supine humans. Anesthesiology 1985; 63: 61623[ISI][Medline]
16 Rooke GA, Freund PR, Jacobson AF. Hemodynamic response and change in organ blood volume during spinal anesthesia in elderly men with cardiac disease. Anesth Analg 1997; 85: 99105[Abstract]
17 Länne T, Lundvall J. Very rapid net transcapillary fluid absorption from skeletal muscle and skin in man during pronounced hypovolaemia circulatory stress. Acta Physiol Scand 1989; 136: 16[Medline]