Original Papers

Vox Sang. 52: 177-181 (1987) 0 1987 S.Karger AG, Basel 0042-9007/87/0523-0177 $2.75/0

The Role of Mean Corpuscular Haemoglobin Concentration in Limiting the Storage Life of Human Blood

Susan Turnera, A . R. Williamsa, J. M. H. Reesb Departments of aMedical Biophysics and bPharmacology, University of Manchester, U K

Abstract. Human erythrocytes stored for more than 21 days in citrate phosphate dextrose with adenine (CPDA-1) at +4 "C had a decreased cell volume, an increased cell density, an elevated mean corpuscular haemoglobin concen- tration (MCHC) and hence an elevated internal viscosity. These changes are normally masked by the reversible cell swelling which accompanies storage in CPDA-1. incubation of stored cells for 24 h at 37 "C in fresh autologous plasma mimics the effects of reinfusion and causes the artificially swollen cells to shrink to their true volume. Thus, incubated cells which had been stored in CPDA-1 for more than 21 days exhibited a decreased filterability through Nuclepore membranes in vitro, which is apparently correlated with a decreased survival time in vivo.

Introduction

The limited storage life of blood inevitably leads to some wastage of donations made to the Blood Transfu- sion Service (BTS). At present, blood may be stored for a period of up to 35 days in citrate phosphate dextrose with adenine (CPDA- I ) , which is the anticoagulant currently used by the BTS. If these time-expired cells were to be reinfused, they would be rapidly eliminated by the body.

Williams and Morris [ 19801 showed that the mean corpuscular haemoglobin concentration (MCHC) of hu- man erythrocytes increased linearly with time in vivo, with approximate values of 3 1.7 g/dl for young cells and 37.5 g/dl for the oldest cells. This increase in MCHC is apparently due to the progressive loss of intracellular ions (and therefore water) without a corresponding loss of haemoglobin. Consequently, the haemoglobin concentra- tion increases as the cell shrinks. Cokelet and Meiselman [ 19681 showed that the viscosity of a solution of haemo- globin increased non-linearly with increasing concentra- tion, so that the internal viscosity of a young erythrocyte would be 5-9 cP, compared with about 50 CP for an old erythrocyte. Williams and Morris [ 19801 proposed that this increase in internal viscosity would delay the transit time of older erythrocytes through the filter network of the spleen and bone marrow and thus increase the prob-

ability of sequestration by macrophages within this filter system.

Young erythrocytes would be expected to traverse the 3 - ~ m wide channels of the spleen in about 0.1 s [Harris and Kellermeyer, 19701, whereas old erythrocytes would take at least 1 s to traverse the same channel. This longer residence time within the splenic filter would increase the probability of contact with a macrophage and hence increase the probability of sequestration of the oldest cells. In support of this hypothesis, it has been observed that any factor which increases the transit time of eryth- rocytes through the filter networks (e.g. a decrease in the surface aredvolume ratio of the cells) shortens their life- span in vivo.

If the time-dependent loss of cell volume which occurs in vivo also occurs as the cell ages during storage in vitro (albeit perhaps at a different rate), the shrunken cells would rapidly be removed on reinfusion because of their high internal viscosities. This process could account for the finite usable life-span of stored blood, even in the absence of significant immunological changes on the sur- face of the cells.

A study was therefore made of the effects of prolonged storage in CPDA-1 on erythrocyte volume, MCHC, cell density and the correlation of these measurements with flow properties of erythrocytes through polycarbonate membranes.

178 Turner/Williams/Rees

Materials and Methods

Experienced clinicians withdrew blood samples from an antecu- bital vein of healthy adult volunteers. The blood was immediately anticoagulated with CPDA-1 in the same ratio as used by the BTS (63 ml CPDA-1 with 450 ml blood) and divided into aliquots. Then, 20-ml samples in sterile Macartney bottles were stored at +4'C. Blood samples were removed from store at weekly intervals up to 42 days after the blood had been taken.

Measurement of the Percentage of Cells More Dense than a Mixture of Phthalic Acid isters (Density- 1.098 g h l ) Blood samples were centrifuged for 5 rnin at 2,000 g in a Centaur

2 MSE centrifuge and the plasma and buffy coat were removed. The packed cells were washed twice using a sterile saline solution (0.9% w/v) and then mixed thoroughly using a Vortex mixer. Glass haema- tocrit tubes (Bilbate) were filled with aliquots of well mixed packed red cell suspensions and sealed at one end using Cristaseal (Hawksley and Sons Ltd.).

A mixture of dimethyl and di-n-butyl phthalic acid esters (Aldrich Chemical Co.) was made, having a specific gravity of 1.098 g/ml, using 0.6 vol dimethyl phthalate (spec. grav.= 1.189 g/ml) and 1 vol di-n-butyl phthalate (spec. grav. = 1.042 g/ml). A 3-mm column of this phthalate-ester mixture was placed on top of each of the packed red cell columns in the haematocrit tubes. The tubes were then spun at an average centrifugal force of 12,000 gfor 30 rnin in an MSE high-speed ultracentrifuge at +4 'C.

A travelling microscope was used to measure the lengths of the columns of packed cells which were above and beneath the ester layer after centrifugation. The length of the column of packed cells beneath the ester layer was then expressed as a percentage of the total length of the red cell column, to give a measure of the percentage of cells more dense than 1.098 g/ml.

Measurement of Mean Cell Volume and MCHC Samples of whole blood stored in CPDA- 1 were used to determine

mean cell volume (MCV) and MCHC using standard blood bank methods with the Coulter S analyser.

Assessment of Red Cell Flexibility using Nuclepore Filtration Samples of whole blood were taken and white cells, platelets and

other debris removed before filtration, so that specific assessment of red cell flexibility could be made. The method used to remove white blood cells and platelets was adapted from a technique used by Lingard [ 19741. A phthalate-ester mixture with a specific gravity below that ofred cells [Danon and Marikovsky, 19641 and above that ofwhite cells [Graham et al., 19551 was used as a barrier medium. The specific gravity of the mixture was 1.09 g/ml and was obtained by mixing 5.0ml of the di-n-butyl with 2.31 ml dimethyl phthalate ester.

1 ml of the ester mixture was placed in a centrifuge tube and 6 ml of the anticoagulated blood were placed on top of the ester mixture. The tube was spun for 30 rnin at 1,900 g. After centrifugation, three distinct layers were seen: a layer of packed red cells at the bottom of the tube, the clear barrier medium containing white blood cells and platelets above this, and a layer of plasma on top [Lingard, 19741. The top two layers of plasma and barrier medium were then removed using a Pasteur pipette. The packed red cells were thoroughly mixed and then diluted with sterile saline (0.9% w/v) to give haematocrits

ranging from 0 to 45%. The method of white blood cell removal was found to be 93.4+0.55% efficient as measured using the Coulter S analyser.

The polycarbonate Nuclepore membrane filters used were 25 mm in diameter and had an average pore size of 5 pm. A 5 y m size was used, since fresh red blood cells passed through readily in our low- pressure system. Expressing the results as flow ratios (c.f. rates) minimises the critical importance of the pore size. Had a 3-lm size been used, higher (non-physiological) pressures would have been necessary and blockage would have occurred more frequently. The diameter of the flow chamber was 12.5 mm, the flow area was 1.26 cm2, the pore density was 4x los cm-2 and the number of pores was approximately 5x los. AU filters were presoaked in sterile saline (0.9% w/v) for at least 24 h before use. This prevented the inclusion of dry pores which would not have allowed passage of red blood cells and would have reduced the effective filter area.

The filtration apparatus was developed from systems described by Reid et al. [I9761 and Schmid-Schonbein et al. [1973]. The Nuclepore filter was positioned in a perspex filtration unit and a 0.4-ml sample of blood was placed on top of the filter. The blood sample was prevented from flowing through the filter under gravity by means of a back pressure of 4 cm H20, obtained from a Hy-flo minipump attached to the filtration unit. The minipump was also used to produce the driving pressure of 2 cm HzO. This was pulsa- tion-free due to the presence of a 5-litre gas reservoir.

The flow of the cell suspension through the Nuclepore membrane caused air to be displaced, which transferred water via a siphon to a collection vessel. The collection vessel was suspended from an iso- metric transducer, so that any change in the weight of water in the vessel resulted in a signal passing to a chart recorder (Rikadenki B-104). After each run, the water surfaces were automatically levelled by the siphoning system.

Non-linearity of traces indicated a blockage of the filter or a leak within the siphoning system, and these results were not used. Saline runs were made before and after each blood run to test for repeata- bility, as a further check that there were no adverse effects. Each filter was only used for one subject, so the system was calibrated before each blood run using sterile saline (0.9% w/v). The haematocrit of each sample was f 1Oh. An accurate estimate of the flow rate through the membrane could therefore be obtained for each sample from the gradient of the weighthime traces.

At each time of sampling, all the methods described were repeated on erythrocytes which had been incubated in autologous fresh plas- ma. Stored red cells were prepared by centrifuging the stored whole blood sample for 5 rnin at 2,000 g in a Centaur 2 MSE centrifuge. The plasma and buffy coat were removed and the red blood cells washed twice using sterile saline (0.9% w/v). Fresh plasma was obtained by centrifuging fresh whole blood for 5 rnin at 2,000 g in a Centaur 2 MSE centrifuge under sterile conditions. The plasma was removed and added to the stored packed red cells to a haematocrit of approx- imately 45%. Samples were incubated at 37 'C for 24 h before use.

Results

The specific gravity of the phthalate-ester mixture (1.098 g/ml) used to measure changes in the percentage of dense cells within a sample was chosen because it had been found to give approximately 5% dense measure-

MCHC Limiting Blood Storage Life 179

Fig. 1. Changes in density (a), MCV (b) and haemoglobin concentration (c) of human red blood cells over 42 days of storage. Density is expressed as percentage of cells more dense than a phthalate-ester mixture (density= 1.098 g/ml). Values are shown for non-incubated stored cells (0) and for stored cells incubated in fresh plasma for 24 h at 37 'C (m). Mean of 6 determinations *SEM. * ~ ~ 0 . 0 5 .

ments for freshly taken cells from a large number of adult donors. The results for the non-incubated cells showed that there was little change in the percentage of cells more dense than the ester mixture over the 42-day storage period. Figure 1 a shows that the percentage of dense cells remained below 10% for non-incubated cells. For cells measured after incubation in fresh plasma, a marked increase in the percentage of dense cells was observed after approximately 2 1 days of storage. This continued to increase with increasing time in store, with over 60% of the red cell population having a density greater than 1.098 g/ml after 42 days of storage. This large increase in the percentage of dense cells within the erythrocyte popula- tion was accompanied by a decrease in the red cell vol- ume, as measured using the Coulter S analyser (fig. 1 b). There was little change in cell volume measured for non- incubated stored cells, but a statistically significant decrease in cell volume of approximately 30fl was observed for stored cells which had been incubated in fresh autologous plasma.

The Coulter S analyser was also used to measure MCHC over the 42-day storage period (fig. lc). Again, there was very little change in MCHC for the non-incu- bated stored samples, but a statistically significant increase in MCHC was found after approximately 2 1 days for stored cells incubated in fresh plasma. The MCHC

was found to be 29.22 f 0.2 g/dl for fresh cells incubated for 24 hand 53.50 k 6.79 g/dl for 42-day stored incubated cells which would result in a marked rise in internal viscosity over the storage period.

As a result of these findings, red cell flexibility mea- surements were made using the Nuclepore filtration method. Freshly taken cells and cells stored for 2 1 days in CPDA- 1 were used because significant changes in red cell volume, MCHC and the percentage of dense cells had been found after this storage time.

Measurements for non-incubated and incubated blood were made, and the results are shown in figure 2.

To eliminate the effects of different filters, the results are expressed as flow ratios and were calculated in the following way:

Rate of flow of red cell suspension Rate of flow of saline

Flow ratio =

These rates of flow were gradients taken from the weighthime plots using the same filter membrane.

Figure 2a shows flow ratios for freshly taken cells before and after incubation for 24 h at 37 "C over a hae- matocrit range of 0-45%. The flow ratio decreased by approximately 7% for a 5% increase in haematocrit. There was no,significant difference between the results obtained for freshly taken cells before and after 24 h of incubation

180 Turner/Williams/Rees

Fig. 2. Flow ratios of fresh human red blood cells (a) and blood cells stored for 2 1 days at 4 'C in CPDA-1 (b) before (0) and after (m) incubation in autolo- gous fresh plasma for 24 h at 37 'C. The broken line indicates that the flow ratio could not be measured for cells above a certain haematocrit. Mean of 6 determi- nations ? SEM. pt0.05.

up to a haematocrit of 45%. It was therefore assumed that the incubation process itself had no effect on red cell flexibility as measured using Nuclepore filtration. Figure 2b shows the flow ratios for 2 1 -day stored cells before and after incubation in fresh plasma. Flow ratios for stored cells after incubation are lower than those for non-incu- bated cells at all haematocrits, with significant differences in flow ratios measured at haematocrits greater than 20% (p-cO.05). Flow ratios of stored cells which had been incubated in fresh plasma could not be measured at hae- matocrits greater than 25% because attempted filtration resulted in filter blockage, so that straight line plots could not be obtained to measure the gradients. The flow ratio for stored cells decreased by approximately 9% for a 5% increase in haematocrit for incubated cells, compared with 6% for non-incubated stored cells.

Discussion

The effect of storage in CPDA-1 on erythrocytes was measured for a period of up to 42 days. There was very little change in MCV, MCHC, percentage of dense cells and red cell deformability over this period for cells which had not been incubated in fresh plasma before these measurements were made. After a 24-hour incubation period at 37 "C in autologous fresh plasma, statistically significant changes were measured for blood stored for periods greater than 21 days. These changes were a decrease in MCV, an increase in MCHC and in the percentage of dense cells within the population, and a decrease in red cell deformability.

This incubation process apparently mimics what would happen to those cells were they to be reinfused in vivo. We can therefore conclude that two distinct pro- cesses are occurring simultaneously while the cells are being stored. One is an irreversible cell shrinkage (which results in an elevated MCHC and a decreased filterability) and the other is a readily reversible cell swelling which is presumably due to the accumulation of CPDA-1 meta- bolites. Incubation for 24 h in fresh plasma (and presum- ably reinfusion in vivo) apparently reverses the effects of this secondary cell swelling phenomenon and results in the production of shrunken cells.

The increase in MCHC which normally occurs in vivo as the cells age [Williams and Moms, 19801 also appeared to occur on storage in CPDA-1 in vitro, although this effect was masked unless the cells were incubated in fresh plasma. This increase in MCHC has been shown to be caused by progressive loss of potassium, without compen- satory sodium gain, resulting in cell dehydration [Clark et al., 19781. Progressive loss of potassium from erythro- cytes is a well documented lesion in CPD-stored blood [Wallas, 19791, and it is known that the intracellular potassium levels do not return to normal for approxi- mately 7 days after transfusion [Crawford and Mollison, 1955; Valeri and Hirsch, 19691.

The high internal viscosity resulting from an elevated MCHC causes the cells to become less deformable [Clark et al., 19781 as was measured using the Nuclepore filtra- tion method. A decrease in the ability to flow through a filter network will increase the probability of erythrocyte removal and provide a possible explanation for the finite storage time for red blood cells.

MCHC Limiting Blood Storage Life 181

It may be possible to reverse this storage lesion and reinflate red blood cells by loading the cells with potas- sium. Reinflated cells may have MCVs, MCHCs, percent- ages of dense cells and flow characteristics similar to those found for freshly taken cells, so this may provide a means of extending the storage time of banked blood.

Lingard, P. S.: Capillary pore rheology of erythrocytes. I. Hydroelas- tic behaviour of human erythrocytes. Microvasc. Res. 8: 53-63 (l 974).

Reid, H. L.; Barnes, A. J.; Lock, P. J.; Dormandy, J. A.; Dormandy, T. L.: A simple method for measuring erythrocyte deformability. J. clin. Path. 29: 855-858 (1976).

Schmid-Schonbein, H.; Weiss, J.; Ludwig, H.: A simple method for measuring red cell deformability in models of the microcircula- tion. Blut 26: 369-379 (1973).

Valeri, C.R.; Hirsch, N.M.: Restoration in vivo of erythrocyte ade- nosine triphosphate, 2,3-diphosphoglycerate, potassium ion, and sodium ion concentrations following the transfusion of acid-

References

Clark, M.R.; Mohandas, N.; Caggiano, V.; Shohet, S.B.: Effects of abnormal cation transport on deformability of desiccytes. J. supramol. Struct. 8: 521-532 (1978).

Cokelet, G. R.; Meiselman, H. J.: Rheological comparison of hemo- globin solutions and erythrocyte suspensions. Science 162: 275- 277 (1968).

Crawford, H.; Mollison, P.L.: Reversal of electrolyte changes in

- citrate-dextrose-stored human red blood cells. J. Lab. clin. Med.

Wallas, C.H.: Sodium and potassium changes in blood bank stored human erythrocytes. Transfusion 19: 2 10-2 1 5 ( 1979).

Williams, A. R.; Morns, D. R.: The internal viscosity of the human erythrocyte may determine its life-span in vivo. Scand. J. Hae- matol. 24: 57-62 (1980).

73: 722-733 (1969).

stored red cells after transfusion. J. Physiol. -129: 639-647 (1955).

Danon, D.; Marikovsky, Y.: Determination of density distribution of red cell population. J. Lab. clin. Med. 64: 668-674 (1 964).

Graham, H. T.; Lowry, 0. H.; Wheelwright, F.; Lenz, M. A.; Parish, H. H.: Distribution of histamine among leucocytes and platelets. Blood 10: 467-481 (1955).

Harris, J. W.; Kellermeyer, R. W.: Theredcell, pp. 281-626(Harvard University Press, Cambridge 1970).

Received May 28, 1985 Accepted: September 5 , 1986

J. M. H. Rees, PhD, Department of Pharmacology, University Medical School, GB-Manchester M 13 9PT (UK)