anesth analg-2003-mcilroy-1572-7

7/29/2019 Anesth Analg-2003-McIlroy-1572-7

1/6

Acute Intravascular Volume Expansion with Rapidly

Administered Crystalloid or Colloid in the Setting of

Moderate Hypovolemia

David R. McIlroy, MB, BS, FANZCA, and Evan D. Kharasch, MD, PhD

Department of Anesthesiology, University of Washington Medical Center, Seattle

Although the distribution of various crystalloid andcolloid solutions at equilibrium has been well estab-lished, the acute peak expansion of intravascular vol-ume that can be achieved with the rapid administra-

tion of crystalloid or colloid is unknown. We studiedeight healthy male subjects in a two-part crossovertrial designed to assess the maximal increase in intra-vascular volume achieved with 1000 mL of lactatedRingers solution compared with the same volume of6% Hetastarch. Subjects were made moderately hy-povolemic by the withdrawal of 900mL of blood, andthen the crystalloid or colloid solution was rapidlyinfused over 57 min. Serial dilution of hematocritwas measured every 5 min for 30 min to determine

changes in blood volume. Peak expansion of intra-vascular volume with lactated Ringers solution was630 127 mL, occurring immediately the rapid infu-sion was complete, whereas the peak expansion of

intravascular volume with 6% Hetastarch was1123116mL and occurred 5 min after the completion of thefluid infusion. The results were significantly differ-ent(P 0.001).Theseresults would suggest that evenfor very short periods of time, rapid infusion of col-loid significantly more effectively increases bloodvolume and, by inference, cardiac output than thesame volume of crystalloid, even if the crystalloid isadministered very rapidly.

(Anesth Analg 2003;96:15727)

IV fluid resuscitation is an integral part of modern

anesthesia practice. The goal is to increase intravas-cular volume as a means to augment cardiac output

and organ perfusion. This may be attempted with acolloid or crystalloid solution. The benefits of each typeof fluid have been widely debated for many years. Al-though the amount of each fluid that remains in theintravascular space at equilibrium has been well docu-mented (1), the ability of a given fluid to acutely andtemporarily expand the intravascular space when givenrapidly to hypovolemic subjects has not been adequatelyinvestigated. Several complex and sophisticated mathe-matical models have predicted the rate of redistribution

of these fluids and thus the increase in blood volume fordiffering rates and duration of crystalloid infusion (2).However, these models have used relatively slow fluidinfusion rates (30 min) and rapidly administered fluid

(100 mL/min) does not fit the constructed pharmaco-

kinetic model (3,4). Other recent studies from the samegroup show that a larger proportion of crystalloid thanexpected is retained in the intravascular space at equi-librium in subjects who are made hypovolemic immedi-ately before the fluid challenge (5,6).

The aim of the present investigation was to comparethe peak increase in intravascular volume achieved withcrystalloid compared with colloid when administered IVrapidly to moderately hypovolemic subjects using a bal-anced crossover trial. The hypothesis was that equalvolumes of crystalloid or colloid would provide thesame initial increase in intravascular volume when ad-

ministered rapidly to moderately hypovolemic subjects.

Materials and MethodsEight healthy males 2236 yr (mean, 29 yr) weighing71103 kg (84 11 kg) were studied. We performed aprospective two-part crossover trial designed to assessthe maximal increase in intravascular volumeachieved with 1000 mL of lactated Ringers solutioncompared with 1000 mL of 6% Hetastarch. Each sub-

ject was studied on two separate occasions at least

Supported, in part, by the Department of Anesthesiology andNIH grant M01-RR-00037 of the University of Washington ClinicalResearch Center.

Accepted for publication January 27, 2003.Address correspondence and reprint requests to David McIlroy,

Department of Anesthesia and Pain Management, Alfred Hospital,Commercial Rd., Melbourne 3004, Victoria, Australia. Addresse-mail to [email protected].

DOI: 10.1213/01.ANE.0000061460.59320.B0

2003 by the International Anesthesia Research Society1572 Anesth Analg 2003;96:15727 0003-2999/03


2/6

6 days apart. Because of the long and slightly uncer-tain elimination half-life of 6% Hetastarch, (7) the or-der of fluids was not randomized, and each subjectreceived the lactated Ringers solution on their firstvisit and then 6% Hetastarch at their second visit. The

study was approved by the University of Washing-tons IRB, and all volunteers gave written informedconsent with the knowledge that they could withdrawfrom the trial at any time. The study was jointlyfunded by an intra-departmental research grant and aNational Institute of Health grant to the University ofWashington Clinical Research Center.

Subjects were admitted to the Clinical ResearchCenter having been instructed to eat a light meal

before their arrival. On arrival, their height andweight were checked and they were asked to void andthen lay down in bed where baseline vital signs wereobtained. A 14-gauge IV catheter was then placed in

each antecubital fossa. Subjects were monitored with3-lead electrocardiography, automated noninvasive

blood pressure measurement, and pulse oximetry andremained supine throughout the procedure. An at-tending anesthesiologist was also present throughoutthe procedure with full resuscitation drugs and equip-ment available. A citrated blood collection bag wasattached to one of the IV cannula, and 900 mL ofvenous blood was drained by gravity over 15 min. Arapid fluid infusing device (Level 1, Smiths IndustriesMedical System, Rockland, MA) was attached to theother IV cannula and was primed with either lactated

Ringers solution (Baxter Healthcare Corporation,Deerfield, IL; electrolyte composition [mEq/L]: Na130, K 4, Ca 2.7, Cl 109, and Lactate 28; 273 mOsm/L)or 6% Hetastarch (Abbott Laboratories, North Chi-cago, IL; electrolyte composition [mEq/L]: Na 154 andCl 154; 308 mOsm/L).

When the withdrawal of 900 mL of blood was com-plete, a 4-mL blood sample was obtained for baselinehematocrit measurement. We then began the rapidinfusion of 1000 mL of fluid via the Level-1 device,which warms the fluid close to body temperature anddelivers it under 300 mm Hg of pressure. LactatedRingers solution 1000 mL was typically delivered in

4 5 min, whereas 1000 mL of 6% Hetastarch took7 8 min to deliver.

The completion of the rapid infusion of fluid wasdesignated time zero (0). A blood sample was ob-tained from the IV cannula not used for fluid infusion.The hematocrit of each 4-mL sample of blood wasmeasured in duplicate (Cell Dyne 3500; Abott, SantaClara, CA) by technicians blinded to which fluid had

been given, and the mean was used for further calcu-lations. Individual hematocrits are accurate to 2 he-matocrit percentage points. Sampling was repeatedevery 5 min for 30 min. At the time of each blood

sample, an initial volume of 5 mL of venous blood was

withdrawn to prevent any contamination of the sam-ple with other fluid. This volume was subsequentlyreturned to the patient. Previous studies indicate thatthe redistribution process of crystalloid is almost com-plete in euvolemic subjects 30 min after completion of

the infusion (4,5).After 30 min, the 900 mL of collected blood wasreturned to the subject over 90 min. Initial blood vol-ume was calculated according to the formula of Na-dler et al. (8):

Blood Volume (L) 0.3669 height3 (m)

0.03219 weight (kg) 0.6041

We then subtracted 900 mL from this value as ourpostphlebotomy baseline blood volume, from whichto calculate the increase in blood volume with fluid

infusion.Changes in blood volume were calculated from se-

rial dilutional changes in hematocrit. Although tradi-tional methods for measuring the amount of fluidremaining in the intravascular space involve indicatordye dilution or dilution of radiolabeled red bloodcells, these methods are unsuitable for studies ofnonsteady-state conditions because they require anunchanged blood volume for a period of 30 40 min to

be accurate (4). Although most of the published workin this area has used hemoglobin dilution to indicatechanges in blood volume (2,4,5,9), the precision of thistechnique has been questioned (10). It has been shown

that the ratio of large-vessel:whole-body hematocritdoes not remain constant in the setting of volumeloading with colloid, and a more precise measure of

blood volume change may be obtained with indepen-dent measures of erythrocyte volume and plasma vol-ume, the sum of which provides blood volume. How-ever, the calculation of plasma volume requires theinjection of indocyanine green, which, because of itsclearance, cannot be repeated more frequently thanevery 20 min. It was therefore deemed unsuitable foruse in this setting, where the peak increase in intra-vascular volume was expected to occur within 5 min

of completion of the fluid infusion, followed by arapid decline in those subjects given crystalloid.

Increase in blood volume (mL)

(baseline hematocrit/measured hematocrit)

baseline blood volume baseline blood volume.

Corrections were not made for the blood samplesremoved (32 mL total) or the lactated Ringers solutionused to flush the sampling line (24 mL total) becausethe small volumes were not felt to significantly affect

the results.

ANESTH ANALG CARDIOVASCULAR ANESTHETIC MCILROY AND KHARASCH 15732003;96:15727 ACUTE INTRAVASCULAR VOLUME EXPANSION: COLLOID VERSUS CRYSTALLOID


3/6

There was no need to include a correction factor forthe standard ratio between large-vessel and whole-

body hematocrit because the initial blood volume cal-culation was not based on a large-vessel hematocritmeasure. Previous studies (11) confirm that 6%

Hetastarch does not significantly alter the rheologicalproperties of blood independent of a simple dilutionaleffect.

The primary outcome measure was the difference inpeak intravascular volume expansion between lac-tated Ringers solution and 6% Hetastarch after com-pletion of the rapid infusion of fluid compared usingStudents t-test. The secondary outcome was time-specific differences in intravascular volume expansion

between lactated Ringers solution and 6% Hetastarchdetermined using two-way repeated-measures analy-ses of variance.

ResultsThe withdrawal of 900 mL of blood from subjects wasgenerally well tolerated. The mean reduction in sys-tolic and diastolic blood pressure was 9 and 4 mm Hg,respectively. The mean change in heart rate was areduction of 9 bpm (range, 29 to 17). One subject

became symptomatically bradycardic at the comple-tion of blood withdrawal with a sinus bradycardia at28 bpm but responded to atropine 1 mg IV and washappy to continue the study. One other subject expe-rienced some mild light headedness that settled with

elevation of the legs. The rapid infusion of fluid via theLevel 1 produced no symptoms.

Rapid infusion of lactated Ringers solution pro-duced no significant change in mean heart rate ormean value of systolic blood pressure. However, themean pulse pressure did increase from 51 12 mmHg to 58 16 mm Hg. By contrast, the rapid infusionof 6% Hetastarch was associated with an increase inmean heart rate from 69 21 bpm to 79 17 bpm.There was also an increase in the mean value of sys-tolic blood pressure from 115 24 mm Hg to 128 19 mm Hg along with an increase in the mean value ofpulse pressure from 51 12 mm Hg to 59 12 mmHg (Fig. 1 and 2).

The measured hematocrit decreased with the rapidinfusion of fluid to values ranging from 29.5 to 39.5and then tended to increase steadily over 30 min in thelactated Ringers group, whereas increasing onlyslightly in the 6% Hetastarch group. Group means forhematocrit and the corresponding calculated increasein intravascular volume ( sd) are shown in Table 1.

The peak increase in intravascular volume with lac-tated Ringers solution occurred immediately aftercompletion of the rapid infusion of fluid (time 0)and was 630 127 mL (Figs. 3 and 4). The peak

increase in intravascular volume with the 6%

Hetastarch actually occurred 5 min after completion ofthe infusion and was 1123 116 mL, which wasslightly larger than the actual volume infused. Thedifference between the colloid and crystalloid wassignificant (P 0.001).

At every time point from 0 to 30 min, there was alarge and statistically significant difference betweenlactated Ringers solution and 6% Hetastarch in theirability to increase the intravascular volume. The dif-ference between the two solutions was smallest imme-diately after completion of the infusion (time 0),593 mL, but grew larger with each 5-min time intervalto 20 min, where the difference in intravascular vol-ume expansion was 708 mL. Beyond 20 min, the dis-crepancy between the two solutions did not increase

further.

Figure 1. Heart rate (bpm) at admission, after 900 mL of venesec-tion, and every 5 min for 30 min after the rapid infusion of eitherlactated Ringers solution or 6% Hetastarch 1000 mL. Time 0 repre-sents immediately when the fluid infusion was complete.

Figure 2. Pulse pressure (mm Hg) at admission, after 900 mL ofvenesection, and every 5 min for 30 min after the rapid infusion ofeither lactated Ringers solution or 6% Hetastarch 1000 mL. Time 0represents immediately when the fluid infusion was complete.

1574 CARDIOVASCULAR ANESTHETIC MCILROY AND KHARASCH ANESTH ANALGACUTE INTRAVASCULAR VOLUME EXPANSION: COLLOID VERSUS CRYSTALLOID 2003;96:15727


4/6

The intravascular volume expanding effect of the6% Hetastarch was well maintained throughout thestudy period. There was no statistically significantreduction in the volume expansion until time 30 min, where the reduction in mean volume expan-sion was a little over 100 mL compared with thevolume expansion at time 510 min.

Not surprisingly, the intravascular volume expand-ing effect of lactated Ringers solution declined rap-idly after its initial peak at time 0. By 15 min aftercompletion of the infusion, the mean increase in intra-vascular volume was only 403 88 mL. Beyond15 min, there was no further significant decline inintravascular volume.

DiscussionThe expansion of intravascular volume by crystalloidand colloid after the process of redistribution has oc-

curred is well documented in physiology texts (1).

However, complete redistribution to equilibriumtakes significant time, and more recent work has fo-

cused on the distribution of these fluids during andshortly after their infusion (25,9). In fact, by examin-ing the early distribution of these fluids, complex andsophisticated mathematical models have been devel-oped that are able to predict the expansion in intra-vascular volume for a given volume of crystalloidinfused over a set period of time. From these models,and the studies on which they are based, it has becomeapparent that IV infused crystalloid is distributed inan expandable space that is significantly smaller thanthe traditional extracellular volume (20% lean bodymass), (2) particularly in the first 30 minutes after

completion of the infusion.

Table 1.

Lactated Ringers 6% Hetastarch

HematocritIncrease in intravascular

volume (mL)a HematocritIncrease in intravascular

volume (mL)a

Baseline 40.1 0 40.5 0Time 0 35.2 572 (113) 32.6 1008 (134)Time 5 35.9 482 (86) 32.5 1018 (110)Time 10 36.2 442 (72) 32.6 1017 (142)Time 15 36.8 366 (77) 32.7 990 (98)Time 20 37 342 (113) 32.8 986 (145)Time 25 37 343 (103) 32.9 957 (150)Time 30 37.4 292 (67) 33.2 917 (149)

a Group means are given in parentheses.

Figure 3. Increase in intravascular volume (mean sd) determinedby hemodilution after the rapid infusion of 1000 mL of lactatedRingers solution or 6% Hetastarch. Time 0 represents the comple-tion of the rapid infusion of fluid.

Figure 4. Peak increase in intravascular volume after rapid infusionof 1000 mL of lactated Ringers solution or 6% Hetastarch. Linesshow the results for individual subjects. Also shown are mean sd.For lactated Ringers solution, the peak increase occurred immedi-ately after completion of the fluid infusion. For 6% Hetastarch, thepeak volume expansion occurred 5 min after completion of theinfusion.



5/6

Hypovolemia also increases the amount of crystal-loid retained in the intravascular space (5,9). Typi-cally, with moderate hypovolemia, blood volumeincreases by approximately 50% of the administeredvolume of crystalloid. Increased volume expansion, at

least temporarily, is also achieved with more rapidinfusions (3,12).Most of the previous studies have infused the crystal-

loid into subjects over a period of 30 minutes. In the onestudy, in which normovolemic volunteers received1500 mL of lactated Ringers solution over 15 minutes,the increase in intravascular volume did not fit the math-ematical models (3). However, the reality of anesthesia isthat moderate or severe hypovolemia occurs often andwhen blood is not instantly available, and much fasterrates of crystalloid or colloid resuscitation must be usedto temporize the situation. In such a situation, does itmatter whether you infuse a colloid or a crystalloid to

these patients while waiting for blood to arrive, so longas you do it quickly?

We have demonstrated a significant difference inthe ability of 6% Hetastarch and lactated Ringerssolution to acutely expand the intravascular volume inmoderately hypovolemic subjects in the period imme-diately after rapid infusion. Despite the moderate hy-povolemia and infusing the fluid rapidly, 1000 mL ofthe lactated Ringers solution could only expand theintravascular volume by 630 127 mL compared withthe 1123 mL 116-mL expansion achieved by 6%Hetastarch under the same conditions. These resultswould suggest that for acute resuscitation, 6%

Hetastarch is significantly more able to expand theintravascular volume and, by inference, the cardiacoutput than is the same volume of lactated Ringerssolution, even in the initial 510 minutes.

The intravascular volume expansion created by thecrystalloid is only slightly larger than that achieved ina previous study where 900 mL of lactated Ringerssolution was infused over 30 minutes to moderatelyhypovolemic volunteers or yet another study where25 mL/kg of crystalloid was administered to normo-volemic subjects over various time periods rangingfrom 15 minutes to 45 minutes. This result is consistentwith the theory that at rapid infusion rates, the com-pliance of the vasculature rapidly becomes less thanthe compliance of a portion of the interstitial spaceand so no further increase in intravascular volumeoccurs.

The only two subjects who experienced symptomatichypotension with the blood withdrawal had the largestincrease in intravascular volume with the lactated Ring-ers solution (744 mL and 798 mL). This may representincreased amounts of autotransfusion in these subjects.Autotransfusion is a powerful homeostatic mechanism.Several studies have addressed this topic and foundvarying results (5,9,13). Some have suggested increases

in intravascular volume of up to 500 mL over 10 minutes

because of autotransfusion, whereas others have docu-mented much smaller volumes. Given that part of themechanism of autotransfusion is believed to be a re-duced capillary hydrostatic pressure operating as part ofStarlings forces, it seems logical that the two patients

who started to decompensate from their hypovolemiamay also have had the greatest degree of autotransfu-sion. It may also suggest that at even more profoundlevels of hypovolemia, the compliance of the vasculaturemay be such that larger amounts of crystalloid can beretained within the intravascular space.

Our observation that the peak expansion of intra-vascular volume with 6% Hetastarch was more thanthe infused volume (1123 mL) and occurred 5 minutesafter the completion of the infusion is not a new find-ing (14). The delayed peak may simply represent im-precision of the measurement technique because be-tween 0 and 25 minutes after completion of the

infusion, there is no statistically significant differencebetween the increased intravascular volume measure-ments. However, the 6% Hetastarch solution isslightly hypertonic with respect to plasma (308mOsm/L), and it may be that this is having a continu-ing effect of increasing intravascular volume in thetime immediately after completion of the infusion.

The finding of a peak volume expansion that islarger than the volume administered may be ex-plained by measurement error, but given that it wasconsistently present throughout all but one of thesubjects, it more likely represents an effect of auto-transfusion. However, given the rapidity with which

autotransfusion can occur, it might be expected that itwould be able to reverse just as quickly with appro-priate restoration of intravascular volume. The in-creased intravascular volume in our subjects receivingcolloid remained more than the 1000 mL infusedthroughout the entire 30-minute measurement period,possibly suggesting that it was caused by the hyper-tonicity of the fluid in addition to a component ofautotransfusion. It may also represent an over estima-tion of the volume effect caused by an alteration of thelarge-vessel:whole-body hematocrit ratio.

Our study may be criticized for having used the he-moglobin dilution method because of the inconstancy ofthe large-vessel:whole-body ratio for hematocrit. Be-cause of the time required for indocyanine green clear-ance (15,16), and the rapid redistribution of lactatedRingers solution after the initial peak increase in intra-vascular volume, the double label measurements of bothplasma volume and erythrocyte volume (10) were notsuitable for our study. However, the magnitude of thechange in large-vessel:whole-body hematocrit ratio witha colloid volume load similar to ours has been demon-strated to be approximately 10%. This is unlikely to havehad a significant impact on our primary outcome.

We may also be criticized for comparing two fluids of

differing tonicity, but we believe that any confounding

1576 CARDIOVASCULAR ANESTHETIC MCILROY AND KHARASCH ANESTH ANALGACUTE INTRAVASCULAR VOLUME EXPANSION: COLLOID VERSUS CRYSTALLOID 2003;96:15727


6/6

effect on the results is likely to have been small. Ideally,we would have used 6% Hextend as our colloid solution,

but it is not available in our institution. Hextend is ahydroxyethyl starch colloid in an electrolyte solutionvery similar to Ringers solution rather than normal sa-

line. We used healthy awake volunteers whose cardio-vascular homeostatic defense mechanisms can beassumed to function optimally. The results of this con-trolled situation of moderate acute hypovolemia may,therefore, not be applicable to an anesthetized surgicalpopulation with more severe hypovolemia and shocksuperimposed on preexisting comorbidities.

Our measurements did not allow us to calculate themagnitude of any autotransfusion. Whereas this maypotentially have affected our raw data for increase inintravascular volume, it will have operated equallyacross both colloid and crystalloid arms of the trialand is unlikely to have significantly affected the cal-

culated difference between the two solutions in theirability to increase the intravascular volume under theconditions studied.

In conclusion, we have demonstrated that undercontrolled conditions of moderate hypovolemia, therapid infusion of lactated Ringers solution 1000 mLover four to five minutes can acutely expand the in-travascular volume by approximately 600 mL. By con-trast, the same volume of 6% Hetastarch can expandintravascular volume by more than 1100 mL. Withinthe limitations discussed, these results would suggestthat even for very short periods of time, rapid infusionof colloid is significantly more able to increase blood

volume and, by inference, cardiac output than is thesame volume of crystalloid even when the crystalloidis given very rapidly.

The authors would like to acknowledge the efforts of Ms ChristineHoffer in the recruitment of subjects for this study.

References1. Guyton AC. Textbook of physiology. 7th ed. London: WB Saun-

ders, 1986;38292.2. Stahle L, Nilsson A, Hahn RG. Modelling the volume of expand-

able body fluid spaces during i. v. fluid therapy. Br J Anaesth1997;78:138 43.

3. Hahn RG, Svensen C. Plasma dilution and the rate of infusion ofRingers solution. Br J Anaesth 1997;79:64 7.

4. Hahn RG, Drobin D, Stahle L. Volume kinetics of Ringerssolution in female volunteers. Br J Anaesth 1997;78:144 8.

5. Drobin D, Hahn RG. Volume kinetics of Ringers solution inhypovolemic volunteers. Anesthesiology 1999;90:8191.

6. Prough D, Svenson C. Current concepts in perioperative fluidmanagement. Anesth Analg 2001;92:70 7.

7. Klotz U, Kroemer H. Clinical pharmacokinetic considerations inthe use of plasma expanders. Clin Pharmacokinet 1987;12:12335.

8. Nadler SB, Hidalgo JU, Bloch T. Prediction of blood volume innormal human adults. Surgery 1962;51:224 32.

9. Riddez L, Hahn RG, Brismar B, et al. Central and regionalhemodynamics during acute hypovolemia and volume substi-

tution in volunteers. Crit Care Med 1997;25:635 40.10. Rehm M, Haller M, Orth V, et al. Changes in blood volume and

hematocrit during acute preoperative volume loading with 5%albumin or 6% hetastarch solutions in patients before radicalhysterectomy. Anesthesiology 2001;95:849 56.

11. Audibert G, Donner M, Lefevre JC, et al. Rheologic effects ofplasma substitutes used for preoperative hemodilution. AnesthAnalg 1994;78:740 5.

12. Ewaldsson CA, Hahn RG. Volume kinetics of Ringers solutionduring induction of spinal and general anaesthesia. Br J Anaesth2001;87:40614.

13. Lundvall J, Lanne T. Large capacity in man for effective plasmavolume control in hypovolemia via fluid transfer from tissue to

blood. Acta Physiol Scand 1989;137:51320.14. Tollofsrud S, Elgjo G, Prough D, et al. The dynamics of vascular

volume and fluid shifts of lactated Ringers solution and hyper-tonic saline-dextran solutions infused in normovolemic sheep.Anesth Analg 2001;93:82331.

15. Iijima T, Aoyagi T, Iwao Y, et al. Cardiac output and circulatingblood volume analysis by pulse dye-densitometry. J Clin Monit1997;13:819.

16. Haruna M, Kumon K, Yahagi N, et al. Blood volume meas-urement at the bedside using ICG pulse spectrophotometry.Anesthesiology 1998;89:1322 8.


anesth analg-2003-mcilroy-1572-7

Documents