Factors Influencing Yield of Plateletpheresis by Discontinuous Centrifugation

H. OGATA, K. NAGASHIMA, N. I INUMA, S. HOSOGAYA, A N D T. AKABANE

From the Blood Bank Section and Central Clinical Laboratory, Shinshu University Hospital, Matsumoto, Nagano-ken, Japan

Platcletphereses performed by the Haemonetics Model 30 Blood Prwessor under a fixed condition were retrospectively analyzed. The mean platelet yield was 5.97 X loll (* 1.46 X loll). Prepheresis platelet count was significantly correlated to platelet yield (p < 0.001). but hematwrit was not (p >0.1). Blood-drawing time showed a significant correlation (p<O.Ol) and may suggat that an infusion rate of 60 ml/minute is a tech- nique of choice.

A large number of platelets (% platelet recovery = 73.0 & 14.1%) was discarded in order to remove eryth- rocytes and leukocytes from the initial product. The recovery was negatively correlated to the percentage removal of leukocytes (p<O.Ol), which means that a platelet product of good quality not containing leuko- cytes can be prepared at the sacrifice of a large quantity of platelets. It is desirable to devise a simple and practical technique to obtain pure platelets without loss.

PLATELETPHERESIS using the blood cell separator is routine in many blood banks. It meets increasing demands for platelet trans- fusions to the patients with bone marrow suppression. The pheresis is particularly useful for obtaining a large amount of plate- lets from a HLA-matched single donor. The Haemonetics Model 30 Blood Processor (Braintree, Massachusetts) has been shown to be a relatively simple and safe means of platelet collection. It has become one of the most widely used instruments, and the tech- nique has been well e~tablished.' '~'" This brief technical note reviews our experience of plateletpheresis with the Haemonetics Model 30 Blood Processor in the search for a quanti- tatively better pheresis procedure.

Received for publication July 15, 1980; accepted August 31, 1980.

Materials and Methods One hundred fifty consecutive plateletphereses

done by the Haemonetics Model 30 Blood Proces- sor were chosen for analysis. All donors were healthy male volunteers, more than 60 kg in body weight. They were selected according to the guide- lines of Japan Red Cross with minor modifica- tions. Complete blood counts and platelet counts were obtained on each donor before and after each pheresis and also on the products. Cells were counted electronically, or manually when necessary.

The procedure of pheresis was essentially the same as previously described."' Blood was pro- cessed eight times through a small bowl (225 ml). Acid-citrate-dextrose formula A was used as anticoagulant. The bowl was usually filled at a rate of 60 to 80 ml/minute. I t was reduced to 20 ml/ minute when platelet-rich fraction was col- lected. The collection was carried out 15 ml beyond the appearance of reddish color in the exit tube. The operator recorded the time of each cycle and condition of blood flow. Actual blood-drawing time was the sum of the time for eight cycles of inflow. It did not include the time for blood return.

The initial products (product I) were centrifuged at 290 X g for six minutes to remove the majority of contaminated erythrocytes and leukocytes. The final product was called product 11. The per cent platelet recovery and the per cent leukocyte removal were defined as follows:

x loo total platelets in product I I total platelets in product I platelet rrcovery =

total leukocytes in I1 ~

total leukocytes in 1 leukocyte removal = [ I -

The YHP-2100 computer was used for statistical analysis.

Results Table 1 lists volume and numbers of plate-

lets and other components in products I and 11. The initial product (product I) contained

0041-1 132/81/ I100/0719 $00.70 @ J. B. Lippincott Co.

Volume 21 719 Number 6 Transfusion November-December 19x1

7 20 OGATA ET AL.

Table 1. Properties of Platelet Products

Transfusion November- December 198 I

n Mean S D

Product I volume ml 150 286.0 31 .O erythrocytes ml 149 38.9 6.7 leukocytes xi 09 150 6.23 1.51 platelets x10” 150 5.97 1.46

Product II volume leukocytes’ platelets

ml 124 232.0 33.0 xi 09 117 0.12 0.20 xl 0” 123 4.43 1.30

Platelet recovery % 123 73.0 14.1

Leukocyte removal 96 117 98.1 3.0

The data deviate markedly from normal distribution

5.97 f 1.46 X 10” platelets (mean f SD). Sev- eral factors were analyzed to obtain better platelet yield in product I, in which prepheresis platelet count and actual blood-drawing time were evi- dently related to the platelet yield (Table 2). The platelet number in product 1 increased as actual blood-drawing time became longer until 110 minutes. With more than 110 minutes of blood- drawing time, there was no more increment. Records of these I5 phereses were inspected and all were found to be associated with difficulties in blood drawing with or without mechanical troubles. Difficulties in blood-drawing were expe- rienced in some phereses that were completed in less than I10 minutes. Therefore smooth phereses with no such difficulties were chosen for a further analysis to find the strength of relationship between platelet yield and actual blood-drawing time. Donor hematocrit was included in the

analysis as a variable, because the volume of blood processed was expected to increase as hematocrit decreased.

Table 3 shows the result of correlation studies of 76 smooth phcreses. The platelcte yield in product I was influenced strongly by the donor platelet count (partial correlation coefficient = 0.8500, p < 0.001). and it also was significantly related to actual blood-drawing time (r = 0.4546, p < 0.01). Donor hematocrit gave a minimal effect on platelet yield.

The process from product I to I1 was studied, showing an average platelet recovery of 123 products to be 73.0% (Table I). There was a tendency that per cent recovery of platelet im- proved according to decrease of per cent leukocyte removal (Fig. I ) . The association was evaluated with Spearman rank correlation coefficient, which gave a significant result (r, = -0.51 16. p C0.002).

Table 2. Platelet Number of Product l

Platelet Count x l(r n Mean SD

Minutes n Mean SD

Relation to Prepheresis Platelet Count C 20 20-25 25-30 30-35 14 26 53 34

4.41 5.16 5.71 6.39 0.70 1.67 0.80 0.88

Relation to Actual Blood-Drawing Time < 70 70-80 80-90 90-100

6 15 31 55 4.72 5.17 5.92 5.98 1.78 0.76 1.16 1.38

35-40 > 40 14 5

7.59 8.62 0.94 0.97

100-110 >110 19 15

6.58 6.11 1.29 1.17

Volume 21 Number 6

recovery %

90

80 -

70

YIELD OF PLATELETPHERESIS 72 1

Table 3. Correlation between Platelet Yield and Three Factors

partial correlation correlation constant coefficient P

Prepheresis platelet count Prepheresis hematocrit Actual blood-drawing time

0.1628 0.0244 0.0371

0.8500 0.0944 0.4546

<0.001 > 0.1 < 0.01

Discussion This study was based on our hope to obtain

better platelet products. It focused solely on quantitative aspects of products of the Hae- monetics Blood Processor.

The mean collection of 5.97 X 10" by a small bowl for eight cycles is comparable to the results of others.47677 It has been reported that platelet yield is directly related to the prepheresis platelet count of the d o n ~ r . ~ ~ * * ~ This factor was confirmed in this study to be most important to an increased platelet yield.

6 0 .

504 n=35 14

r = -.5116 S

3s 3 3

leukocyte -98.5 removal %

98.5- 99.0- 99.5-

FIG. I . Platelet recovery and leukocyte removal.

The factor is usually and practically uncon- trollable. Although less strongly related, actual blood-drawing time is another factor with statistically significant correlation to platelet yield. Unless the long time was the result of difficulties complicating blood draw- ing, a larger number of platelets was har- vested at slower inflow rate (p<O.OI). The exact cause of this phenomenon is unknown. A better separation of platelet fraction with a longer centrifugation may be one of the related factors. The amount of blood pro- cessed is another explanation of better yield, as a longer centrifugation packs erythrocytes more tightly and permits a larger volume of inflow. The latter explanation appears to be a minor factor, because the donor hematocrit, which influences the volume of blood pro- cessed, gave almost no, or only a minimal effect on platelet yield (p > 0.1). Whatever the cause, a slower rate of blood drawing might be better to obtain a larger number of platelets: 60 ml/minute instead of the usual rate of 60 to 80 ml/minute. If a slower rate was adopted, we would have to pay more attention to functional aspects of platelet. It has been reported that the in virro functions of platelets prepared by Haemonetics Model 30 were altered?" The platelet fraction con- taining erythrocytes and leukocytes was markedly affected, and the pure fraction showed mild changes.* Although the exact cause of impaired function of platelets has not been established, a longer centrifugation of a slower rate of inflow is likely to produce an adverse effect on platelets. This effect may be improved by using an anticoagulant with a low concentration of citrate?I2

OGATA ET AL. 722 Transfusion November- December I98 I

There is another often overlooked but even more influential process of platelet collection. I t is the centrifugation to removeerythrocytes and leukocytes from the initial product. A large number of platelets are lost in this process. A mean loss of 15 to 23 per cent has been rep~r ted .”~” Our result of 27.0 per cent seemed to suggest an inadequate centrifugal procedure. However, a close observation dis- closed that platelet recovery was negatively related to leukocyte removal. A 91.7 per cent removal of leukocytes was associated with 23 per cent of platelet recovery: and our average leukocyte removal was 98.1 per cent. A better recovery can be achieved with toler- ating contamination of a slightly larger number of leukocytes in the final product. This is controlled partially by expressing the platelet-rich plasma from the original bag to the satellite. Although a larger con- tamination is a shortcut to avoid waste of collected platelets, those leukocytes are clini- cally unde~irable.~ For the present, a better leukocyte removal would be an appropriate choice in spite of the poorer platelet recovery, since a minimal number of leukocytes in a platelet product has not been established.

The method of platelet collection using the Haemonetics Model 30 Blood Processor has been well established, and there is little way to improve platelet yield. It may be more important for blood banks to devise a simple preparation technique to obtain pure platelet from the initial product with minimal loss.

References I . Aisner J. Schiffer CA. Wolff J H , Wiernik PH. A

standardized technique for efficient platelet and leukocyte collection using the Model 30 Blood Processor. Transfusion 1976;16:437.

2. Chao FC, Tullis JL. Tinch RJ. Conneely GS, Baudanza P. Plateletpheresis by discontinuous centrifugation: effect of collecting methods on the in virro function of platelets. Br J Haematol 1978;39: 177.

3. Herzig RH, Herzig GP, Bull MI. er a / . Correction of poor platelet transfusion rrsponses with leukocyte-poor H L-A-matched platelet concen- trates. Blood 1975;46:743.

4. Huestis DW, Fletcher JL. White RF, Price MJ. Citrate anticoagulants for plateletpheresis. Trans- fusion 1977; 17: 151.

5. Mishler J M , Borberg H, Reuter H, Gross R. The utilization of a new strength citrate anticoagulant during centrifugal plateletpheresis. 11. Assessment of in virro platelet function. Blut 1977;34:237.

6. Nusbacher J , Scher JL, MacPherson JL. Platelet- pheresis using the Haemonetics Model 30 Cell Separator. Vox Sang 1977;33:9.

7. Reiss RF, Katz AJ. Statewide support of thrombo- cytopenic patients with ABO matched single donor platelets. Transfusion 1976;16:312.

8. Schiffer CA, Aisner J, Wolff JH. Platelet and granulocyte collection using the Model 30 Blood Processor. Proc Adv Pheresis Seminar. Hae- monetics Research Institute, 1975.

9. Szymanski 10, Patti K, Kliman A. Efficacy of the Latham Blood Processor to perform platelet- pheresis. Transfusion 1973; 13:405.

10. Ts’ao CH, Wirman JA. Ruder EA. Altered in virro functions of platelets prepared by the Haemo- netics Blood Processor. J Lab Clin Med 1975; 86:315.

1 I . Tullis JL, Tinch RJ, Baudanza P. er a/ . Platelet- pheresis in a disposable system. Transfusion 1971;11:368.

12. Zitko M, Reuter H, Borberg H, Mishler JM. The utilization of a new strength citrate anticoagulant during centrifugal plateletpheresis. 111. Assess- ment of in virro platelet metabolism. Vox Sang 1979:36:347.