INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

CHANGES IN HEMATOCRIT (Hct) and/or hemoglobin (Hb) concentration in blood are widely used for estimating relative changes in plasma volume (PV) during acute changes in posture (1, 7). The validity, however, of this method for estimating relative changes in PV during these circumstances has not yet been adequately documented by comparison with a direct tracer-dilution method. We have previously demonstrated that during seated, thermoneutral water immersion (WI) to the neck in humans, the use of changes in Hct and Hb according to the method of Dill and Costill (2) underestimates the changes in PV by >50% compared with estimations by an Evans blue (EB) dye dilution technique developed for repeated measurements (13, 14).

The difference between the two methods for calculating PV changes during WI might be caused by differences in the F cell ratio [the ratio of whole body Hct to large-vessel Hct (10, 16), where whole body Hct is determined by simultaneously measuring PV and red cell volume] for the following reasons. During immersion, a change in local Starling forces induced by the pressure of the water column on the tissues (14, 17) facilitates filtration of fluid into the capillaries, thus increasing PV. This might induce a relatively larger decrease in Hct in the peripheral vessels (arterioles, capillaries, and venules) compared with that of the more central vessels (i.e., the vena cava or a cubital vein), thus introducing a possible significant change in F-cell ratio. Consequently, an error will occur in the calculation of relative PV changes from changes in Hct and Hb.

Thus, on the basis of these theoretical considerations and on previous results (14), we hypothesized that changes in Hct and Hb lead to underestimation of the relative changes in PV during an acute posture change from seated to 6° head-down tilt (HDT). Such an underestimation might be caused by the same mechanism supposed to occur during immersion, i.e., by changes in Starling forces in the capillaries of the legs. The seated position was chosen as the reference, because PV in humans remains stable in this position for at least 13 h (14).

The relative changes in PV estimated from the changes in Hct and Hb were compared with those of simultaneous measurements by an improved EB method. The improvement of the EB method consisted of multiple postinjection samplings (6) to take into account the possible confounding effects of changes in the transcapillary escape rate of albumin-bound EB due to changes in local Starling forces (9). Furthermore, to avoid the artifactual effects of local changes in hydrostatic forces on blood sampled from a dependent vs. a nondependent arm during the posture change (3), we took the following precautions. 1) Both arms were kept at heart level on all occasions, and 2) blood was sampled from a central venous catheter.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Ten healthy human males [age 23.7 ± 0.5 yr, weight 75 ± 2 kg, and height 1.83 ± 0.02 m] participated in the experiment. All had a negative history of cardiovascular or kidney diseases and exhibited normal values of a routine clinical examination, including arterial pressure measurements and electrocardiogram (ECG) recordings. None of the subjects took any medication at the time of the study. After careful oral and written explanation, written consent was obtained and the experimental protocol was approved by the Ethics Commitee of Copenhagen (100.1698/90) and was in compliance with the Declaration of Helsinki.

No food or fluid intake was allowed during the 12 h before the experiment. From 10 PM the evening before the experiment, the subject was confined to the laboratory. He was awakened at 7:30 AM. While the subject was under local anesthesia and continuous ECG monitoring, a polyethylene central venous catheter (Secalon Kathy 1, Viggo-Spectramed) was led through a cubital vein into the intrathoracic region for measurements of central venous pressure (CVP) and collection of blood. Intrathoracic placement of the central venous catheter was confirmed by typical CVP waveforms and responses to respiratory maneuvers. In the opposite arm, a short 18-gauge venous catheter (Venflon, Viggo-Spectramed) was inserted into a forearm vein for injection of EB. Thereafter, the subject emptied his bladder, drank 400 ml of water, and was weighed.

The subject was thereafter placed in the upright, seated position, with the feet resting on the floor, upper body and lower legs vertical, and thighs horizontal, for 1.5 h. He was subsequently placed in the HDT position on a tilt table for 1 h and was finally placed again in the upright, seated position, also for 1 h (recovery). During all procedures, both arms were placed exactly at heart level (on supports when subject was seated). Room temperature varied maximally as means of the 10 experiments between 24.4 ± 0.2 and 25.8 ± 0.2°C. Arterial pressures (in duplicate), CVP, left atrial diameter (LAD), and heart rate (HR) were determined at 15- or 30-min intervals. At 15- or 30-min intervals, blood was drawn from the central venous catheter. Furthermore, seven samples for determination of EB in plasma were collected on three occasions before, during, and after HDT. A total of 150 ml of blood was collected from each subject. The subject voided at hourly intervals after measurements had been performed and drank 200 ml immediately afterwards. The procedure was always performed in the following sequence: sampling of blood, determination of arterial pressures, CVP, HR, LAD, and finally determination of arterial pressures in repeat.

Blood samples for measurement of plasma concentration of protein (P_prot), Hb, plasma colloid osmotic pressure (COP), Hct, and plasma density (PD) were immediately transferred to 10-ml polyethylene tubes containing 15 IU of heparin/ml of blood.

PV was determined 1 h and 15 min after having the subject seated upright during control, 0.5 h after initiation of HDT, and finally 1 h after termination of HDT, also with the subject seated. EB [2.5 ml (5 mg/ml)] was injected into the bloodstream through the peripheral venous catheter. The syringe was thereafter flushed with 15-20 ml of isotonic saline to wash out residual dye from the syringe and catheter. A baseline value in triplicate (3 × 3 ml) was drawn 1 min before injection. Five, seven, ten, and fifteen minutes after injection, 3 ml of blood were collected from the central venous catheter and transferred to polyethylene tubes containing 15 IU of heparin/ml of blood. The exact points in time of sampling were recorded. After centrifugation for 20 min at 1,500 g, concentration of dye in fresh plasma was measured in a spectrophotometer, and PV was calculated as described in detail previously (6, 14). The coefficient of variation for the EB method (including sampling and pipetting errors, etc.) was 6.1 ± 1.0% (5 repeated PV determinations in n = 6 subjects).

Theoretically, adequate mixing of dye might not be complete for the first two EB samples in the upright, seated position because of slower blood flow in the dependent regions. Therefore, the two initial values of EB might be too high, thereby producing incorrectly low estimates of PV. To explore this issue, we recalculated PV using only the 10- and 15-min blood samples obtained from the subjects in the upright, seated posture. This revealed no difference compared with using all four samples between the fifth and fifteenth minute (3,205 ± 142 vs. 3,260 ± 149 ml in the seated position before HDT and 3,219 ± 181 vs. 3,140 ± 211 ml in the seated position after HDT). Therefore, mixing of dye in the upright, seated position is complete within the initial 7 min.

Hct was measured in quadruplicate by centrifugation of microhematocrit tubes (Brand) in a centrifuge (Microfuge, Christ) for 5 min at 12,600 g. Hct values were corrected for trapped plasma by factoring raw values with 0.96 (8). Hb concentration in blood was measured in duplicate by a spectrophotometric method, as described previously (14). The coefficients of variation of the Hct and Hb measurement methods (including sampling and pipetting errors, etc.) were 1.0 ± 0.1 and 1.2 ± 0.1%, respectively (10 repeated measurements in n = 8 subjects).

P_prot was measured in duplicate on fresh samples in a refractometer (Pocket Refractometer, Bellingham and Stanley). PD was determined in a densitometer (DMA 46, Paar). Changes in PD are an indirect measure of changes in protein concentration in plasma (11). Plasma COP was measured in duplicate in a colloid osmometer (4400 colloid osmometer; Wescor, Logan, UT) on cooled samples.

Hct and Hb values were used to estimate relative changes in PV (%PV) by Eq. 1 (2, 10), using the control value 15 min before initiation of HDT as the baseline

(1)

In addition, PD, Hct (11), and P_prot were used to calculate relative changes in PV by Eqs. 2, 2a, and 3, respectively

(2)

where

(2a)

and

(3)

The subscript "seat" refers to the pre-HDT value, subscript "hdt" refers to the HDT value, and FD is the density of shifted fluid. Hct values in Eq. 2a are given as ratios.

CVP, arterial pressure, and HR were measured as previously described (12). The reference level for the CVP measurements was chosen at the fourth intercostal space when the subject was seated and at the mid-axillary line during HDT.

LAD was measured by M-mode echocardiography during the end-expiratory phase of respiration (Aloka SSD 500, Simonsen & Weel) from standard images obtained from the parasternal long axis view and was averaged over 6 to 12 heart beats.

HR was calculated as the mean over a minimum of 30 s from continous ECG recordings.

Systolic and diastolic arterial pressures (SAP and DAP, respectively) and arterial pulse pressure (PP) were measured with automatic oscillometric equipment (Propaq 102; Dameca, Rødovre, Denmark) on the upper arm. During measurements, the upper arm was hanging alongside the body when the subject was seated so that the cuff was exactly at heart level.

Data are presented as means ± SE. A multifactorial analysis of variance (ANOVA) (Statgraphics 5.0) for repeated measures, with the variable as the main variate and time and subjects as factors, was used to evaluate the effects on a variable over time. To compare relative changes in PV of the different methods at the same experimental points in time, an ANOVA for repeated measures was used, with the variable as the main variate and method (EB, Hct/Hb, P_prot, PD/Hct) and subjects as factors. Differences between mean values were evaluated by a post hoc multiple-range test (Newman-Keuls). Logarithmic transformation of data was performed before analysis if heterogeneity of variances was observed. Paired two-sided t-tests were applied when appropriate. A significance level of 0.05 was chosen.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Changes in PV. PV (Fig. 1) increased within 30 min of HDT by 293 ± 60 ml from 3,205 ± 142 ml (P < 0.05), amounting to a relative increase of 9.3 ± 2.0%. One hour after termination of HDT, PV returned to baseline (3,219 ± 181 ml).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Mean relative changes in plasma volume (PV) before, during, and after 6° head-down tilt (HDT), derived from measurements of Evans blue (EB) dye dilution () and changes in hematocrit and hemoglobin concentration (Hct/Hb, ). Values are means ± SE of 10 subjects. # Significantly different from all values (P < 0.05) before HDT; * significantly different comparing EB with Hct/Hb measurements at similar experimental point in time (P < 0.05).

Cardiovascular variables. CVP increased immediately upon initiation of HDT (P < 0.05) from between 2.8 ± 0.4 and 2.3 ± 0.5 mmHg to between 6.2 ± 0.7 and 6.9 ± 0.8 mmHg. LAD (n = 9) increased promptly from between 16 ± 2 and 17 ± 1 mm to between 26 ± 1 and 27 ± 1 mm (P < 0.05) during HDT. Neither SAP, DAP, PP, nor mean arterial pressure (n = 9) changed during HDT. HR decreased promptly and remained low during the entire HDT period (54 ± 2 to 57 ± 3 beats/min during HDT vs. 62 ± 3 to 66 ± 3 beats/min before HDT, P < 0.05).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The results of this study clearly demonstrate that changes in Hct and Hb underestimate the relative change in PV by as much as 50% during an acute change in posture in humans from seated to HDT. This is in compliance with previous observations from our laboratory that changes in PV during WI in humans are underestimated by the Hct/Hb method (13, 14). Therefore, it is absolutely essential during antiorthostatic maneuvers such as WI and posture changes to utilize a direct (e.g., dye dilution) method for correct estimations of changes in PV.

For the Hct/Hb method to be valid during an experiment, several preconditions have to be fulfilled. 1) The total number of circulating erythrocytes must be unchanged, 2) the sampling site (i.e., a cubital vein) must remain at same level relative to the heart throughout an experiment to avoid the confounding effects of changes in local hydrostatic gradients, and 3) the F cell ratio must remain unchanged.

Regarding points 1 and 2 above, these preconditions seem to have been fulfilled in the present study. First, the amount of blood (20 ml) sampled between the seated prevalue and the value at 30 min of HDT was very small. Thus any error that might occur as a consequence of this sampling lies within the measurement error. Second, the arms remained at same level relative to the heart throughout the whole experiment. Therefore, the underestimation of PV by the Hct/Hb method is not caused by changes in the amount of circulating erythrocytes or local hydrostatic gradients.

Regarding the third point, a possible mechanism for the underestimation of changes in PV by the Hct/Hb method during changes in posture and WI is that the F cell ratio decreases during anti-orthostatic posture changes. Induced by decreased intravascular hydrostatic gradients, increased filtration of fluid into the capillaries of the legs occurs during antiorthostasis. Due to plasma skimming effects and differences in erythrocyte and plasma circulation times (16), this will in turn possibly induce a relatively larger decrease in the Hct of the peripheral, small vessels compared with that of central, large vessels. This will decrease the F cell ratio and thus invalidate the use of the Hct/Hb method for estimating relative changes in PV. It was not the purpose of the present study, however, to directly address this issue, and the suggested mechanism therefore remains hypothetical.

This is the first time that the change in PV in humans has been determined with a direct dye-dilution method comparing the seated position with HDT. It has previously been demonstrated with a direct dye-dilution method in normal subjects and goiter patients that PV decreases by 325 ± 33 ml (12%) within 30 min of a posture change from supine to standing (18). Thus the decrease in PV going from the horizontal to the standing position is comparable in magnitude to the decrease in PV when going from HDT to the seated position.

Our results, however, do not agree with those of Fawcett and Wynn (4) who observed no discrepancy when comparing changes in PV estimated from EB dilution with changes in Hct within 1 h of a posture change from supine to standing. In the study of Fawcett and Wynn (4), the subjects were allowed to walk in the upright postinjection sampling period, and blood was collected from a peripheral venous catheter. It was not stated whether the sampling site remained constant relative to heart level during this experiment. Thus the combined effects of body movements and blood sampled from a peripheral, dependent vein may explain the discrepancy between their results and ours.

The advantage of the EB method used in this study is that several postinjection samplings are made. Therefore, the effects of changes in transcapillary escape rate of albumin-bound EB induced by the different hydrostatic gradients during orthostasis and antiorthostasis are taken into account. Furthermore, due to the repeated injection of dye for each PV determination and the sampling of preinjection baseline samples, changes in intravascular protein content should have no effect on the PV calculations.

It might be argued, however, that mixing of dye in the circulation is not complete within the initial 7 min in the seated position. Thus the EB method might overestimate the change in PV going from the seated to the HDT position. This argument, however, is refuted by our demonstration of similar PV values compared with the original values using only the 10- and 15-min blood samples for calculating PV in the seated position. Thus mixing of dye in the seated position is complete within the initial 7 min.

The relative change in PV estimated from changes in Hct and Hb in this study is less than that observed by some other investigators (1, 7) who sampled blood from peripheral veins. However, it was not clearly stated whether the arms were kept at the same level relative to the heart during all of these experiments. Because an error can be introduced by sampling blood from a vein in a dependent arm (3), we have avoided this problem by sampling blood from a central venous catheter and by having both of the subjects' arms placed at heart level during the entire course of the study.

In contrast to the observations of the present study, we have demonstrated that during intravascular volume changes induced by saline infusion (12) and during 6° head-down bed rest (15), Hct and Hb may confidently be used for determining relative changes in PV, provided that the body position during these interventions remains virtually unchanged compared with a control so that no changes in orthostatic stress and thus no changes in F cell ratio are induced.

The results of this study clearly demonstrate that in humans changes in Hct and Hb lead to an underestimation of changes in PV by ~50% during a physiological change in posture. The underestimation of PV by the Hct/Hb method might be due to changes in the ratio of whole body to large-vessel Hct induced by the changes in orthostatic stress. This, however, needs to be further investigated.

Perspectives