in mammals, the ability of the spleen to eject its contents into the active circulation during times of stress can enhance oxygen transport and, therefore, improve aerobic performance (21, 28, 30). The role of the spleen in sequestering ∼50% of the total erythrocyte volume in seals (15) and horses (28), during times of inactivity, also dramatically reduces the viscosity of the blood and, therefore, the work of the heart. In comparison, the human spleen contains only 10% of the total erythrocytes and has been primarily thought of as a lymphoid or defense organ (26).

Observations of a reduction in spleen size after psychological and physiological stresses have been reported in humans (1, 7, 9, 10,12, 16, 19, 23). However, quantification of in vivo spleen volume is difficult. The anatomical position of the spleen means that isolated views, without overlapping organs, are often impossible. The majority of research, involving technetium labeling of erythrocytes, has utilized only a posterior scan of the spleen and reported relative radioactivity (1, 7, 9, 10, 19, 23). Scintagraphric spleen volumes have been reported only once in the literature after a diving-related intervention (7).

The physiological significance of the human spleen releasing erythrocytes has recently been raised (19, 24), not for its role in enhancing oxygen transport but for its effect on indirect measures of plasma volume. Changes in plasma volume are generally calculated indirectly from hemoglobin and hematocrit. The standard techniques of Evans blue dye or radioisotope-labeled human serum albumin (I-RISA) are prohibitive by their invasive nature and cost. Indirect assessments have become the norm and are presently accepted throughout the world. In these frequently used equations (4,29), the underlying assumption is that circulating erythrocyte volume is consistent with total erythrocyte volume. The ability of the spleen to release up to 10% of the total erythrocyte volume into the peripheral circulation has been overlooked. A recent investigation has suggested that the spleen could account for 25% of the increase in hematocrit (19). This would introduce substantial errors into the indirect assessment methods of plasma volume changes.

Therefore, the purpose of the present study was to quantify splenic erythrocyte volume after submaximal and maximal exercise and determine whether the increased circulating erythrocyte volume confounds calculated plasma volume.

METHODS

Subjects.

Nine healthy male subjects (age 25 ± 1.0 yr, height 176 ± 3.0 cm, mass 75 ± 2.8 kg, and peak oxygen consumption 54 ± 2.3 ml · kg⁻¹ · min⁻¹) completed the study. Before any testing, the subjects received a verbal description of the experiment and were required to complete a written consent for this study, which was approved by the University of British Columbia Ethics Committee.

Experimental protocol.

Subjects were required to report on two separate occasions. The first day of testing comprised a maximal oxygen consumption test on an electronically braked cycle ergometer, conducted at the Allan McGavin Sports Medicine Centre on the University of British Columbia campus, as described previously (27).

The second testing session was conducted in the Nuclear Medicine Department of Vancouver General Hospital. Subjects were asked to report to the laboratory in the morning after an overnight fast and to refrain from exercise for 24–36 h before the study. After the subject had been supine for 30 min, an intravenous catheter was inserted into both the right and left antecubital veins, and baseline levels of all hematological variables were determined.

The results of the peak oxygen consumption test from day 1 determined the initial power output at which each subject began exercising. A workload corresponding to 60% was selected. This was designed to be below each individual's anaerobic threshold, and 10 min cycling at comparable intensities had previously been shown to produce significant reductions in spleen radioactivity (19). The exercise session consisted of work bouts of 5-, 10-, and 15-min duration, with 30-min rest between each workload (Fig.1). The order of the workloads was randomized. Steady state was maintained by monitoring oxygen consumption (K4b², Cosmed, Rome, Italy) throughout each work bout and adjusting the power output accordingly. A count of spleen activity was conducted immediately preceding and following each exercise bout.

After adequate time for the subject to recover from the last exercise workload, an incremental exercise test to exhaustion was performed by using an identical protocol as that performed on the first day of testing. At the end of the incremental exercise test to exhaustion, blood samples were drawn in conjunction with spleen activity counts at 0, 10, 20, 30, and 40 min postexercise (Fig. 1).

Spleen volume.

Five milliliters of red blood cells were labeled with the use of 1.5 mg of stannous pyrophosphate followed by 25 mCi of ^99mTc pertechnetate. Labeling yield of red blood cells with the use of this method is >95% (22). Anterior and posterior views of the spleen were taken by using an ultra-high-resolution collimator of a gamma camera (MultiSpect 2, Siemens Medical Systems). The image was counted for 2 min by using a 20% energy window around^99mTc. To determine spleen counts for the anterior and posterior views, a region of interest was drawn around the posterior image, which contained no overlap from surrounding organs, and this was duplicated onto the anterior image (Fig.2). A geometric mean was obtained by averaging the anterior and posterior decay-corrected activity counts. An intraclass coefficient of 0.92 was calculated for the nuclear medicine-determined regions of interest.

A 1-ml sample of red blood cells was obtained during each of the spleen activity counts after the 5-, 10-, and 15-min rides and at 40 and 50 min after the incremental ride to exhaustion. The decay-corrected count of activity in the known volume (1 ml) was used to determine an elution factor. The spleen volume was then calculated from the decay-corrected geometric mean spleen activity counts of the anterior and posterior images at each of the measurement times by using the elution factor.

Limitations.

The anatomical position of the spleen means that isolated views, without overlapping organs, is often impossible. Anterior and posterior views were chosen in the present study after a comparison with oblique views revealed no significant differences. Radioactivity was captured simultaneously over a 2-min period from the anterior and posterior views. Volume calculations are problematic, as a third dimension, depth, is impossible to obtain due to other visceral organs being included in the lateral view. The weighting of the anterior and posterior activity counts evenly is erroneous, as the spleen is predominantly a posterior organ. However, without a lateral view and the intersubject variability in spleen dimensions, the ability to correctly weight each subject's anterior and posterior counts separately is lost.

Spleen size.

A two-dimensional area of the spleen was calculated from a region of interest drawn around the spleen from the posterior view. A pixel count of the area selected was achieved, with part pixels being counted as whole. A pixel represented an area of 0.24 cm².

Blood volume determination.

To determine red cell volume, 15 ml of blood were withdrawn into a syringe containing 3 ml of acid citrate dextrose solution. This blood was then added to 150 μCi NaCr⁵¹. After 20-min incubation, 50 mg ascorbic acid were added, and the solution was allowed to incubate at room temperature for 3 min. Another 10 ml of venous blood were then obtained for measurement of background radioactivity, and 10 ml of the acid citrate dextrose-blood-NaCr⁵¹-ascorbic acid solution were reinjected into the subject. Blood samples were drawn from the opposite arm 30 min postinjection and immediately after the cessation of exercise. Red cell volume was calculated on the basis of the dilution of the reinjected labeled red blood cells (17).

Plasma volume was determined by measuring the dilution of the injected¹²⁵I-RISA (17). Blood samples were drawn 10, 20, and 30 min after the injection. The net counts per minute of these samples were plotted on a semilogarithmic scale and extrapolated totime 0. This extrapolated time 0 count value was used to calculate plasma volume based on the relative dilution of the original injected label (25). Postexercise plasma volume was determined by using the same procedure with ¹³¹I-RISA. Total blood volume was calculated as the sum of plasma volume and red cell volume. Plasma, red cell, and total blood volumes were expressed in absolute terms (liters) and after normalization for body weight (ml/kg).

The change in plasma volume (ΔPV) from pre- (Pre) to postexercise (Post) was also calculated via the following equations: first, hemoglobin and hematocrit (4)

Hemoglobin and hematocrit.

Before exercise and immediately after the incremental ride to exhaustion, a 5-ml blood sample was collected into sterile vacutainer tubes containing EDTA and stored at room temperature until analysis. A blood count was determined by using standard techniques on a Coulter counter at the Department of Pathology and Laboratory Medicine, Vancouver Hospital and Health Sciences Centre, British Columbia.

Plasma viscosity.

At the same time periods as the standard blood counts were determined, a 7-ml blood sample was drawn. This sample was centrifuged for 10 min at 3,000 rpm (Sorvall GLC-2, Kendro Laboratory Products, Newton, CT). With the use of a pipette, 0.95-ml samples of plasma were removed and transferred into a clean 7-ml test tube, and the mass of the plasma was determined (AB104, Mettler-Toledo, Columbus, OH). The average mass of the 0.95-ml plasma samples was recorded, and the density of the plasma was calculated. The plasma sample was then transferred to a viscometer (Cannon-Manning Semi-Micro Viscometer, Cannon Instruments, Philadelphia, PA) and placed in a constant-temperature water bath (37°C) maintained by a hot plate (Thermix Stirrer model 120 MR, Fisher Scientific, Nepean, ON). The viscosity of the sample was determined by measuring the time required for the sample to pass between two fixed points on the viscometer and comparing this to the time required for a sample of water, with a known density and viscosity, to pass between the same two points.

Plasma catecholamines.

Plasma epinephrine and norepinephrine were measured in each of the blood samples drawn (preexercise, immediately after the incremental ride to exhaustion, and every 10 min, for 40 min). Two 5-ml samples were collected into precooled EDTA blood collection tubes. Samples were immediately placed on ice and processed within 0.5 h of collection. Blood samples were spun at 3,000 rpm for 15 min in a refrigerated centrifuge. The plasma was then transferred (a minimum of 2.5 ml) into two separate aliquots, which were immediately stored at −70°C until analysis. Plasma catecholamines were assayed by HPLC (HPLC 1090, Hewlett Packard, Andover, MA) by using electrochemical detection (LC-4B, Bioanalytical Systems, West Lafayette, IN) and commercially available reagents (reagent kit, 195-6074, Bio-Rad Laboratories, Mississauga, ON). The technique, which has been documented previously (11), has a minimum detection limit of 0.05 and 0.15 nmol/l for epinephrine and norepinephrine, respectively, and a coefficient of variation of <3% (personal communication with Head of the Division of Clinical Chemistry, Vancouver Hospital and Health Sciences Centre, British Columbia).

Statistics.

Mean values and measures of variability were determined for anthropometric and descriptive maximal oxygen consumption data obtained during preliminary screening. Data from the 5-, 10-, and 15-min rides were compared with a 3 (exercise duration) by 2 (time) two-way factorial analysis of variance with repeated measures on both factors. Recovery data from the incremental ride were compared with a one-way analysis of variance with repeated measures. When significantF-ratios were observed, Scheffé's test was applied post hoc to determine where the differences occurred. Pearson product-moment correlations were applied to the semilog plots of the plasma catecholamine concentrations and spleen volume data. Hematological parameters were compared by using t-tests for dependent samples (pre- vs. postexercise). Plasma volume reduction was compared with a one-way ANOVA. The level of significance was set atP < 0.01 for all ANOVA and correlation procedures. Statistical power calculations were performed a priori to estimate an appropriate minimum sample size. A sample size of eight was calculated.

RESULTS

Submaximal exercise.

The effect of varying duration but maintaining exercise intensity on spleen volume is reported in Table 1 and Fig. 2. There was no significant effect of the different durations on spleen volume. A significant time effect was observed in spleen volume (preexercise = 85.8 ± 10.1 ml, postexercise = 59.3 ± 8.1 ml; P = 0.0006), indicating a pre- to postexercise change across all durations.

Oxygen consumption, heart rate, and power outputs before and during the different duration exercise periods are displayed in Table 1. There was no significant effect in preexercise heart rate or oxygen consumption. Nor was there any significant difference during the three exercise periods of different duration for power output, heart rate, or oxygen consumption. There was a significant increase in heart rate (preexercise = 69 ± 3.6, average exercise = 144 ± 3.9 beats/min; P < 0.0001) and oxygen consumption (preexercise = 5.6 ± 0.4, average exercise = 32.4 ± 1.4 ml · kg⁻¹ · min⁻¹;P < 0.0001) when preexercise was compared with the average exercise values across all exercise durations.

Maximal exercise.

The effect of recovery from maximal exercise on spleen volume is reported in Table 2. After the maximal ride, there was a significant time effect observed (P< 0.0001). Immediately after (0 min) and 10 min after maximal exercise (10 min), spleen volumes were significantly reduced compared with preexercise (rest = 85.9 ± 9.93, 0 min = 36.4 ± 3.95, 10 min = 61.9 ± 8.51 ml; P < 0.0001). There was no significant difference between rest and 20, 30, and 40 min. These effects were identical when spleen volume reduction was represented as a percent change from rest (Table 2). Spleen size and volume were positively correlated (r = 0.89,P < 0.0001). Statistical significance was achieved for spleen size at identical comparisons as with spleen volume (Table 2).

Plasma norepinephrine and epinephrine levels are also represented in Table 2. There was a significant time effect for both norepinephrine (rest = 1.51 ± 0.30, 0 min = 29.09 ± 6.84 nmol/l;P < 0.0001) and epinephrine (rest = 0.13 ± 0.05, 0 min = 3.17 ± 0.82 nmol/l; P < 0.0001). There was no significant difference between rest and 10, 20, 30, and 40 min for norepinephrine or epinephrine. Spleen volume during recovery from maximal exercise was inversely correlated with both plasma catecholamine concentrations {spleen volume (%rest) = 7.8 − 19.2 ln[norepinephrine]; r = 0.84,P < 0.0001; spleen volume (%rest) = 36.7 − 15.35 ln[epinephrine]; r = 0.70, P < 0.0001}.

Blood volumes from before and immediately after the exercise session are displayed in Table 3. Plasma volume showed a significant reduction after exercise when represented as a function of body mass (preexercise = 46.6 ± 1.44, postexercise = 38.3 ± 1.83 ml/kg, P = 0.002), as well as an absolute value (P = 0.026; Table3). The reduction in plasma volume was not significantly correlated with the change in hematocrit pre- to postexercise (n = 9, r = −0.50, P = 0.17). There was no significant change in red cell volume or total blood volume, although, when expressed relative to body mass, total blood volume approached significance (preexercise = 77.3 ± 2.85, postexercise = 68.5 ± 3.14 ml/kg, P = 0.054). Figure3 shows the calculated reduction in plasma volume after exercise using equations based on hemoglobin and hematocrit values compared with radioisotope labeling. There was no significant difference between any of these methods in calculating plasma volume changes.

Table 3 illustrates the changes pre- to postexercise in standard blood count variables. There was a significant increase in hemoglobin, hematocrit, white and red blood cell counts, and platelet count after exercise. Hematocrit values were significantly correlated with spleen volume (n = 18, r = −0.49,P = 0.038), but the exercise-induced hemoconcentration was not related to absolute (n = 9, r = −0.04, P = 0.91) or relative changes (n = 9, r = −0.18, P = 0.64) in spleen volume. Plasma viscosity also significantly increased after exercise. Only the mean cell volume and red cell distribution showed no significant change from preexercise values.

DISCUSSION

In vivo observations of the response of the human spleen to exercise are rare (1, 9, 10, 19, 23). The present investigation is the first to quantify the spleen response to submaximal exercise of different durations and during recovery from a maximal exercise challenge. The consistent decrease in spleen size and volume after different durations of exercise suggests that the spleen's response is intensity dependent. The small variation in spleen volume measurements after maximal exercise and the semilogarithmic correlation with plasma catecholamines suggest that the spleen is actively contracting to a minimal size. The present investigation was also the first to quantify the effect of the splenic release of erythrocytes on indirect measures of plasma volume. The release of a statistically insignificant amount of erythrocytes by the spleen had no effect on circulating erythrocyte volume and, therefore, peripheral hematocrit. This resulted in the indirect calculations of plasma volume being equivalent to radioisotope-labeled measurements.

Splenic contraction.

The present study is the first to measure the change in splenic volume in response to submaximal exercise of different durations. No differences were observed in the reduction in spleen size and volume among the 5-, 10-, and 15-min exercise bouts when the intensity was maintained constant (Table 1 and Fig. 3), supporting the theory that the spleen is contracting due to an intensity-dependent signal. Without any active constriction of the internal splenic vasculature and assuming the level of flow through the splenic artery is related to the level of sympathetic radioactivity, the average transit time of an erythrocyte through the spleen should be constant. If the mechanism for the spleen size and volume reduction were a passive collapse, then a linear relationship with the duration of the exercise would have been expected. As this was not the case, an active constriction of the splenic vasculature must have occurred in response to an intensity-dependent mechanism.

No differences were observed between resting spleen size and volumes, heart rates, or oxygen consumption before each exercise bout (Table 1). This indicates that the spleen had adequate time to recover from the previous exercise period and that the resting status of the subject was also not compromised by the prior exercise. No differences were observed in the average oxygen consumption, heart rate, or power output during each of the different durations of exercise protocols (Table 1). Therefore, the authors are confident that the intensity of exercise was identical over the three different exercise durations.

A 44 and 56% decrease in spleen size and volume, respectively, were observed immediately after the incremental ride to exhaustion (Table2). Decrements in splenic radioactivity ranging from 39 to 54% have been recorded at exhaustion after graded and continuous maximal exercise protocols (1, 9, 10, 23). The variation in the activity changes could be explained by methodological limitations. The use of the preexercise region of interest for the postexercise scan, despite observations of a reduced spleen size and the possibility of the spleen moving within the abdominal cavity, could affect the activity count. As the degree of movement of the spleen observed in the present study was minimal, it is possible to assume that the spleen stayed within the preexercise region of interest in the previous studies. However, movement of other organs, i.e., the heart and kidneys, may affect the region of interest and artificially inflate the spleen activity. In the present study, the selection of a new region of interest for each image accurately portrays the change in spleen size associated with exercise.

The similar change in spleen radioactivity after maximal exercise in the different studies (1, 9, 10, 23) and the small variation in maximal exercise volumes compared with other measurement periods, achieved in the present study, are all suggestive of the spleen having a limit to its minimal size. This could be indicative of the spleen maximally contracting.

This is the first study to determine spleen size and volume during recovery from an incremental cycle to exhaustion. The spleen returned to preexercise size and volume after 20 min of supine rest (Table 2). Allsop et al. (1) observed splenic radioactivity for 40 min after supramaximal exercise and indicated that the spleen had not fully recovered in this time period. The small sample size (n = 5) of that study and the large intrasubject variability in radionucleotide measurements could account for the conflicting results.

The present investigation is the largest, to date, to measure spleen volume changes after exercise and vasoactive hormones in vivo in humans. The decrease in spleen size and volume in the present investigation correlated with increasing plasma catecholamine concentrations in a semilogarithmic fashion. The only other study to measure vasoactive hormones in conjunction with spleen imaging discovered a gradual, nonlinear decrease in spleen erythrocyte content with increasing plasma catecholamine concentrations (19). The small sample size (n = 5) prevented any statistical treatment, but the authors did propose that the change in spleen size was partially caused by the increased adrenergic activity. Justification for a causal relationship is warranted when comparisons are made to animal studies where infusion of catecholamine induced splenic contraction and increased circulating red cell mass (14,20). The identification of α-adrenoreceptors in the human splenic vasculature (2) and the role that these have been shown to have in contraction of the seal (3) and dog (20) spleen in response to catecholamine stimulation indicate that a similar mechanism may be responsible for the human exercise response.

Additional mechanisms undoubtedly assist in regulating splenic size and volume. The variations of plasma catecholamine concentrations that we measured do not completely account for the changes in spleen size and volume. The ability of the human splenic capsule to assist in contraction has often been overlooked, mainly due to a scarcity of smooth muscle. However, small volume changes have been recorded after stimulation of the splenic nerve (2), indicating that, in humans, direct neural innervation aids in the reduction in spleen size and volume. The passive collapse of the spleen, in response to splanchnic vasoconstriction, may have a small effect on the decrease in spleen size and volume. Attributing a significant role would be unwise, as the evidence from the present study argues against a passive collapse. However, it could represent a minor contributing factor, in combination with stimulation of the splenic nerve, during the initial stages of the reduction in spleen volume when plasma catecholamine concentrations remain near basal levels.

Splenic erythrocyte effect on hemoconcentration.

Before exercise, the splenic red cell pool represented 3.8% (range 2–6.9%) of the total circulating red blood cells and, after exercise, only 1.6% (range 0.7–2.7%) (Table 3). Whereas the relative change in spleen volume is consistent with previous findings, the absolute volumes are significantly smaller. Espersen and colleges (7) reported resting spleen volumes of 193–371 ml, whereas Laub et al. (19) estimated that 200 ml could have been released from the spleen during maximal exercise. The differences arise from the methodologies employed in calculating the spleen volumes. Laub et al. measured splenic radioactivity and, from the percent change in radioactivity, calculated a volume from an assumed “normal” 300-ml spleen. Espersen and colleges (7) used a similar methodology as the present study, with the activity of a known volume of red blood cells being compared with the activity of the spleen. They, however, utilized only a posterior scan and determined depth by measuring an attenuation factor in only two subjects. The lower volume calculated in the present investigation is possibly due to the spleen region of interest being drawn with removal of any influence of overlapping organs, i.e., heart and kidneys, and the mean of the simultaneous anterior and posterior scans being used to determine depth.

The reduction in plasma volume (Table 3) was within the normal range for graded continuous exercise (5). The fluid shift from the intra- to the extravascular compartment caused an increase in plasma viscosity due to an increased concentration of plasma proteins (Table 3). Although not measured directly, whole blood viscosity would also have been expected to increase.

The role of the spleen as a storage vehicle for red blood cells has been hypothesized as a way for the body to regulate the viscosity of the blood. As a non-Newtonian fluid, whole blood viscosity depends on flow rate (18). Therefore, in the seal and horse, as the spleen is capable of containing up to 50% of the red cell volume, the impact on viscosity of a maintained elevation in circulating red cell volume during periods of resting flow rates would place extreme loads on the heart (6, 8). While the human spleen is capable of storing significantly less red blood cells, the role of reducing resting blood viscosity may still be an important cardioprotective function of the human spleen.

The maximal exercise challenge resulted in a 13.6% increase in hematocrit (Table 3). Historically, the increase in exercising hematocrit has been fully attributed to a decrease in plasma volume and, as such, has been used as a way to calculate plasma volume changes. No difference in plasma volume changes, after 60 min of recovery from supramaximal exercise, has been reported between estimations using hemoglobin and hematocrit changes and direct measurement with radiolabeled albumin (1). However, the use of only ¹²⁵I-RISA for both pre- and postexercise measures would have resulted in an overestimation of the reduction in plasma volume, as leakage of albumin from the extravascular space is accelerated during exercise. For accurate quantification, another isotope (¹³¹I-RISA was used in the present study) is required to measure postexercise plasma volume.

The red blood cells that were released from the spleen during exercise did not affect the measurement of plasma volume (Fig.4). The small size and volume of the spleens in the present study were not responsible for the lack of an effect on circulating hematocrit, as the relative change in spleen volume was not correlated with the increase in hematocrit. Interestingly, the change in plasma volume could not account for the change in hematocrit either, although the small sample size undoubtedly affects the lack of a statistically significant correlation.

In summary, the release of red blood cells from the spleen had no significant effect on circulating red cell volume and, therefore, peripheral hematocrit. The role of the spleen in storing red blood cells during periods of inactivity or low stress may be important in reducing the viscosity of the blood; it does not, however, appear to have a role in the hemoconcentration observed after acute, intense exercise.

FOOTNOTES

• Present address of I. B. Stewart: School of Human Movement Studies, Queensland University of Technology, Brisbane, Queensland 4059, Australia.

• Address for reprint requests and other correspondence: I. B. Stewart, School of Human Movement Studies, Queensland Univ. of Technology, QLD 4059, Australia (E-mail:i.[email protected]edu.au).

• The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

• First published December 13, 2002;10.1152/japplphysiol.00040.2002