British J

O U ~ Aof~

Haematology, 1976, 34, 377.

Clinical Experience with Transfusion of Cryopreserved Platelets CHARLES A. SCHIFFER, JOSEPH AISNER AND PETER H. WIERNIK

Clinical Oncology Branch, National Cancer Institute, Baltimore Cancer Research Center, Baltimore, Maryland, U.S.A. (Received 24 February 1976; acceptedfor publication 6 March 1976) SUMMARY. Multiple units of platelet concentrate obtained by plateletpheresis of normal, 'random' or HL-A matched donors were pooled and frozen in polyolefin bags using 5% dimethylsulphoxide (DMSO) as a cryoprotective agent and a controlled freezing rate of I"C/min. The platelets were stored at approximately - 120°C for as long as 201 days, thawed rapidly at 37°C' washed once and resuspended in ACD plasma prior to transfusion. Two different final concentrations of platelets ( 2.7 and 9.0 x 10' '11.) were studied. Twenty-three thrombocytopenic patients have received a total of 40 frozen platelet transfusions. The mean freeze-thaw loss was 21% and was similar for both platelet concentrations. All transfusions were well tolerated and there were no side effects attributable to the small amounts of DMSO infused. Increments in platelet counts I h after transfusion ranged from o to 102 x 10~11. with an overall mean corrected increase in evaluable patients of 12 800 (increasex surface area (m2)/number of platelets transfused x 10"). Corrected increases tended to be greater with the low concentration of platelets. Overall, the increase in count for the frozen platelet transfusions was 65% of the increments obtained with fresh platelet transfusions administered within I week of the frozen platelets. Bleeding times were partially corrected after four out of six transfusions with post-transfusion counts greater than 5 0 x 1o9/I., and active haemorrhage was controlled in some patients by frozen platelet transfusions. These results indicate that pooled platelets can be frozen, thawed and transfused with reasonable efficiency. The frozen platelets can circulate and function haemostatically and may eventually play an important role in supportive care. Platelet transfusion has become an integral feature of the management of patients with leukaemia and aplastic anaemia and has been a major factor in the increased survival now achieved in these patients. The demand for platelets has increased considerably in the last decade and will undoubtedly increase further as more aggressive therapy is administered to patients with other types of tumours. Recently, multi-unit plateletpheresis techniques have been described (Szymanski et a2, 1973 ; Schiffer et al, 1974) which would theoretically decrease the number of donors supplying platelets to each patient thereby decreasing the risk of alloinimunization and hepatitis transmission. Investigations dealing with platelet cryo22

Correspondence:Dr Charles A. Schiffer, Baltimore Cancer Research Center, University of Maryland Hospital, South Greene Street, Baltimore, Md 21201, U S A .

3 77

378

C. A. Schiyer,]. Aisner and P. H. Wiernik

preservation have also intensified in recent years and have as their goal the creation of storage ‘banks’ of frozen platelets from single donors which could provide a constant source of platelets for emergencies as well as HGA matched platelets. It has been shown that single units of platelets from normal donors can be frozen using dimethylsulphoxide (DMSO) as a cryopreservative, and can circulate normally and correct aspirin-induced abnormalities in the bleeding time after autologous retransfusion (Handin & Valeri, 1972; Valeri, 1974). There are few studies in thrombocytopenic patients. The most encouraging results have been reported by Kim & Baldini (1974) and by Slichter & Harker (1972) who demonstrated that platelets frozen as single units with DMSO can circulate and function haemostatically in thrombocytopenic recipients. If large-scale platelet cryopreservation is to be practical, it must be demonstrated that multiple units of platelets can be processed, frozen and transfused together. We have previously reported that multiple units of autologous platelets obtained from patients with leukaeniia in remission can be frozen, retransfused during relapse of leukaemia, and can be effective in the prevention and treatment of haemorrhage in these patients (Schiffer et a!, 1976b). It was found that freeze-thaw recovery was maximal when polyolefin bags (see also Kim& Baldini, 1973),a freezing rate of I°C/min, and lower final concentrations of platelets were used. In vivo platelet recovery, as assessed by posttransfusion increments in counts, also appeared to be better at lower platelet concentrations, although increments were usually inferior to results obtained with fresh platelets. In the present study, multiple units of platelets from normal donors were frozen together using modifications of these techniques in an attempt to assess the therapeutic efficacy of heterologous frozen platelets and to develop a practical method for freezing large numbers of platelets for clinical use. MATERIALS AND METHODS Platelets for freezing were obtained from normal donors using standard serial centrifugation plateletpheresis techniques (Schiffer et al, 1974) and acid-citrate-dextrose (ACD) anticoagulant. In most instances, two units were obtained by plateletpheresis of unmatched, ‘random’ donors, and ABO-identical units were pooled and processed for freezing. For seven transfusions, multiple units of platelet concentrate (PC) from HL-A matched donors were obtained using either serial centrifugation methods or the Model 30 Blood Processor (Haemonetics Corporation, Natick, Massachusetts) as described previously (Aisner et al, 1976). The PC were allowed to stand at room temperature for approximately I 11 to eliminate platelet clumping prior to preparation for freezing. Two different techniques were employed to test the importance of the final concentration of platelets being frozen. In the ‘high concentration’ method, 8-10 units of PC were pooled and centrifuged at 5000 g for 6 min at 25°C. The supernatant plasma was then extracted and the platelets were resuspended in a volume of 30 ml. The platelets were then transferred to 100 ml polyolefin freezing bags (Hemoflex 1000-2, Union Carbide, Chicago, Illinois) to which 300 nil of 10% DMSO in autologous plasma was added over 15-20 min. In the ‘lower concentration’ group four units of PC, or approximately 3 x 10” platelets, were pooled, concentrated by centrifugation,resuspended in a final volume of 50 ml and transferred to a 200 ml polyolefin freezing bag (Hemoflex 2030-2, Union Carbide) to which 50 nil of 10% DMSO in

Frozen Platelet Transfusion 379 autologous plasma was slowly added. The final DMSO concentration was therefore 5% in both groups. During the addition of the DMSO-plasma the platelets were constantly agitated to assure complete mixing. Considerable heat is generated upon the addition of DMSO to plasma and the DMSO-plasma was allowed to cool to room temperature prior to mixing with the platelets. In addition, any possible osmotic injury to the platelets was eliminated by the slow addition of DMSO-plasma to the platelets (Schiffer et a!, 1976a). The final bags were then immediately placed in protective metal containers and frozen to - 80°Cat a rate of I"C/min using a controlled rate freezing apparatus (Kalmbach & Mardiney, 1972). With this apparatus, the rate of freezing is precisely monitored and regulated by a thermocouple inserted into a test tube containing an aliquot of the 5% DMSO-platelet mixture. Extra liquid nitrogen is automatically added as the mixture passes through the heat of fusion, thereby decreasing the rise in temperature which would otherwise take place. The frozen platelets were stored in the vapour of a liquid nitrogen freezer until use. Platelet thawing and transjiision. Platelets were thawed by immersion in a 37°C water bath for 4-5 min and transferred to a polyvinylchloride plastic bag (TA-2, Fenwal, Morton Grove, Illinois). Plasma, I I O ml, obtained and frozen at the time of plateletpheresis, and 10 ml of ACD, which served to reduce platelet clumping (Handin & Valeri, 1972)~were added over 10-15 min. The slow addition was designed to minimize osmotic stress to the platelets. The platelets were then centrifuged at 5000 g for 5 min at 25°C and resuspended in approximately IOO ml of autologous plasma. If insufficient autologous plasma was available, ABO identical ACD plasma was used. Both the freezing and thawing procedures require approximately 20 min of technician's time if automatic pumping equipment is used. Transfusions were administered through standard blood filters over I 5-45 min within 1-2 h of thawing. The recipients were patients with leukaemia or other malignancies who Whenever were profoundly thrombocytopenic (plateletcounts were usually below zo x 10~11.). possible, the transfusions were given as 'prophylactic' transfusions to stable, afebrile, noninfected patients in order to study the results of transfusion under the most favourable clinical conditions. Platelet counts were obtained pre-transfusion, I h post transfusion and daily thereafter. Absolute increments in counts were corrected for the number of platelets administered and the patient's body surface area (BSA) using the formula: Corrected count increment (increment x m2/10" platelets) = (Post-transfusion count- Pre-transfusion count) x BSA (m2)/Io1' platelets infused As an illustration, one would expect a corrected count increase of approximately 1 8 x 1o9/I. after transfusion of I x 10" platelets to an 'average' 1.76 m2 adult with a blood volume of 5 1. assuming that one-third of the platelets are initially sequestered in the spleen (Aster & Jandl, 1964). In some patients bleeding times were performed 2 h post-transfusion using the standardized template technique (Mielke et al, 1969). To assess freeze-thaw preparation loss, platelet counts were determined on the final bag prior to transfusion and compared to platelet counts on a prefreeze specimen obtained before the addition of the DMSO. All platelet counts were done electronically (Bull et al, 1965) in triplicate using a Coulter Thrombocounter (Coulter Electronics, Hialeah, Florida). After two transfusions serial blood samples were obtained and platelet aggregation was measured using the method of Born & Cross (1963) with adenosine 5'-diphosphate (ADP, Sigma Chemical, St Louis, Missouri) as the aggregating agent.

380

C.A. Schifer, J. Aisner and P,H. Wiernik RESULTS

Forty transfusions have been administered to 23 patients. A total of 52 frozen platelet bags have been thawed because in some instances two bags were pooled and transfused together. The pH of the transfused platelets ranged between 6.9 and 7.3 except for one transfusion in which the pH was 6.6. As shown in Table I, the mean overall platelet loss during freezing and thawing was 21%. There was a somewhat lower platelet loss from the higher platelet concentration although the difference from the low concentration was not statistically significant. There was a three-fold difference in platelet concentration between the two groups. Assuming that a ‘unit’ of PC contains approximately 7.5 x 10” platelets, the low concentration bags contained approximately I ‘unit’/a8 in1 while the high concentration bags averaged I ‘unit’/ g nil. TABLE I. Freeze-thaw platelet loss

All transfusions (n = 52) High concentration (n = 15) Low concentration (n = 32)

Per cent loss

Pre-jieeze platelet concentration ( x 109/l.)

21.of2.2* (0-66) 16.7k2.6 (0-42) 22.6k2.7 (0-66)

9 . 1 k 0 . 4 106* ~ (5.3-11.1) 2.7k0.1 x 106 (1.9-3.2)

-

The difference in per cent loss between the high and low concentration groups is not statistically significant. * Mean f SEM, ranges in parentheses.

When examined by phase microscopy, approximately 50-60% of the thawed platelets retained a discoid shape, 30-40% were spherical and the remainder were fragmented or had dendritic projections. Significant clumping was noted in only one preparation. The platelets were stored for an average of 63 d (range 3-201 d) prior to thawing and transfusion. There was no correlation between the duration of storage and either the freeze thaw loss or the clinical results.

Clinical Observations An average of 4.4 x 10” platelcts (range 2.0-7.1 x 10”) or the equivalent of approximately 5.7 ‘units’ of platelets were administered with each transfusion. The transfusions were well tolerated by the recipients and there were no local or systemic side effects attributable to the small amounts of DMSO infused. Three transfusions to patients with massive splenomegaly and three transfusions to alloimmunized patients refractory to random donor platelets were excluded from evaluation. In one patient post-transfusion counts were not obtained. The remaining 3 3 transfusions were considered to be clinically evaluable and the corrected count increments from these transfusions are shown in Table 11. Absolute count increments ranged between o and 102x 109/l. and the mean overall corrected count increment was 12.8 x 10~11. Although the mean corrected count increments were similar in both concentration groups,

Frozen Platelet Transjiusion

381

TABLE 11. Post-transfusion platelet count increments

Corrected count increment at I h ( x 10~11.) No. of transfusions

AII transfusions

Low concentration

High concentration

12.85 1.4* (0-29.5)

1z.g+1.7(0-26)

11.7k2.8(0-29.5)

33

23

I0

* Mean+ SEM with ranges in parentheses.

the patients in each of the groups were not precisely comparable. All of the 10 evaluable highconcentration transfusions were from pooled random donors whereas seven of the lowconcentration transfusions were from HL-A matched donors and 16 were from random donors. More importantly, a higher proportion (6/ 10) of the high-concentration transfusions were given to splenectomized patients as compared to the low-concentration group (6/23). It is likely that this imbalance would tend to give the impression of increased recovery in the high-concentration group. One splenectoinized patient received a total of nine random donor frozen platelet transfusions thereby allowing comparison between the two different concentrations in the same patient. The mean corrected increment was greater in the low concentration group (mean = 22.3 & 1.8 (SEM), range 17-23.3 x 109/l., n = 4) than in the high-concentration group (14.2) 5, range 0-29.5 x 1o9/I., n = 5 ) although the highest increment resulted after a high-concentration transfusion. The clinical course of this patient is summarized in Fig I ; the increments in counts and rate of fall of the platelet count were similar after both frozen and fresh platelet transfusions. Fifteen patients received 16 transfusions of fresh random donor or HL-A matched platelets within I week of receiving frozen platelets and had either I h or 18 h post-transfusion platelet counts obtained which could be compared with the frozen results. The mean corrected count , increment for the frozen platelet transfusions was 65% (k9.5 (SEM), range o - I ~ o n~ =~ 16) of the increments obtained with fresh platelet transfusions. Retransfusion because of persistent thrombocytopenia was usually required every 2-3 days, an interval similar to that following fresh platelet transfusions. Most of the frozen platelets were administered as ‘prophylactic’ transfusions and therefore it was difficult to determine precisely the haemostatic effectiveness of the frozen cells. Only one patient required retransfusion because of continued haemorrhage and in this patient gross haematuria was not controlled by transfusion of fresh platelets. Following three transfusions, epistaxis (two patients) and mild uterine bleeding (one patient) were controlled by frozen platelets and in another patient a transtracheal aspiration was performed without complication at a post transfusion platelet count of 42 x 10~11. Template bleeding times were done 1-2 h after six transfusions which elevated the platelet count to above 50 x 10’11. Pre-transfusion bleeding times were not done because of concern about the risk of skin infections in these patients who were usually concurrently granulocytopenic. In our experience, however, the bleeding time is always greater than 2s min with platelet counts below 25 x 10~11.Bleeding times were reduced to 14 min or less following

C.A. Schi$er,J. Aisner and P.H. Weirnik

DAY

FIG I. The arrows indicate the dates of platelet transfusion and, where known, the number of platelets transfused. Frozen platelets are indicated in broken lines and fresh platelets in solid lines. T.P. was a splenectomized, non-alloimmunized patient with non-lymphocytic leukaemia in whom similar results were achieved with fresh and frozen platelets.

prior to transfusion

I h post transfusion; bleeding time :18min

18 h; bleeding time: llmin I

0

I I

I

I

2

3

I 4

I

5

Minutes

FIG 2. Random donor frozen platelets. Improved aggregation and shortened bleeding times were noted with in vivo circulation.

three ofthe transfusions with post-transfusion platelet counts above 85 x 109/l., to 18 minafter one transfusion at a count of 95 x 1o9/I., but remained above 20 min in two patients whose platelet counts were 55-60 x 10~11.Bleeding times and post-transfusion platelet aggregation studies were performed serially on two occasions. The results of both studies were similar and one of the serial studies is described in Fig 2. There was progressive improvement in ADP induced aggregation after in vivo circulation. In addition, the bleeding times also improved, although in the example shown in Fig z the platelet count fell from 95 to 58 x 10~11. during

Frozen Platelet Tran$usion

383

the interval between tests. These observations suggest that some degree of platelet rejuvenation or repair can occur following transfusion. In six experiments prior to transfusion, the previously frozen platelets did not aggregate even when concentrations of ADP as high as 10- M were used. Although freeze-thaw injury undoubtedly plays the major role in producing this defect, it is possible that the presence of residual amounts of DMSO also served to inhibit aggregation (Schiffer et al, 1976a). Additional washes and resuspension in autologous ACD plasma did not, however, improve aggregation in three experiments. DISCUSSION The present study demonstrates that multiple units of platelets can be processed and frozen together with an acceptable freeze-thaw loss and can circulate and function haemostatically after transfusion to thrombocytopenic recipients. The in uiuo recovery was approximately two-thirds of that seen with fresh platelets. Although not specifically compared, it is likely that haemostatic function, as assessed by bleeding times, was not comparable to ‘normal’, fresh platelets. Nonetheless, active haemorrhage was treated successfully in most of the patients who were bleeding at the time of transfusion. It is of interest that uterine bleeding was controlled in one patient despite the fact that the bleeding time remained above 20 min suggesting that, at least in some patients, the bleeding time may not provide accurate assessment of in uiuo haemostatic function. Successful transfusions in terms of in uiuo recovery and haemostatic effectiveness were seen using both platelet concentrations, although it appeared that the results were more consistent with the lower concentration of approximately one ‘unit’/z~-30ml. Because ‘batches’of four units represent a convenient ‘dose’ to collect, and transfuse together, we plan to use this technique and concentration in future investigations. It is not clear, however, that the concentration of platelets plays a critical role and it may be that higher concentrations are practical and desirable. It is important that the cross-sectional area of the freezing bag (and presumably the heat exchange) be kept constant whatever final volume and concentration are investigated. In the present study, the bag thickness was approximately 1.5-2.5 mm and was similar in both concentration groups. A number of variables in the freezing process could potentially have contributed to platelet injury and the poorer count increments achieved with some of the transfusions. Previous investigations have suggested the importance of slow addition (Schiffer et al, 1976a) and removal of DMSO (Kim & Baldini, 1974), polyolefin freezing bags (Kim & Baldini, 1973 ; Schiffer et al, 1976b), minimal storage time after thawing (Kim et a2, 1974) andafinalDMS0 concentration of approximately 5% (Valeri, 1974; Murphy et al, 1974). These techniques, therefore, were employed in the present study. It is unlikely that the cryoprotective agent itself significantIy affected the platelets and the results of transfusion. Previous in uitro studies have shown that exposure of platelets to 5% DMSO does not affect morphology, size, platelet factor-3 availability and non-stimulated nucleotide or serotonin release (Schiffer et a/, 1976a). Platelet aggregation and thrombininduced release is inhibited in a dose-related fashion but this adverse affect is completely reversible when the platelets are washed and the DMSO removed. Murphy et a1 (1974) demonstrated that preincubation of fresh platelets with 5% DMSO prior to transfusion did

3 84

C. A. Schiyer, J. Aisner and P. H. Wiernik

not produce significant changes in platelet half-life or per cent recovery although decreased recovery was noted with 10 and 15% DMSO. Finally, the small amount of infused DMSO was well tolerated by all the recipients and was not associated with the nausea, vomiting and local venospasm which was described by Djerassi et a1 (1966) who utilized a technique in which the DMSO was not washed out prior to transfusion. The effect of freezing rate on platelet viability has not been adequately studied. Slow rates (I°C/min) which do not attempt to decrease the heat of fusion (Kim & Baldini, 1974), controlled rates (I-S"C/min) in which extra coolant is added during the heat of fusion (Schiffer et al, 1976b),rapid rates of cooling up to the heat of fusion followed by controlled cooling rates (Murphy et al, 1974) and simply placing the bag in a low temperature freezer (Valeri et al, 1974) have all been used with variable success. Unfortunately, there is no easily reproducible in vitro test which will predict post-transfusion platelet function and the ultimate evaluation in all of these studies depended on measurement of post-transfusion recovery. This is obviously a somewhat cumbersome measurement and has prevented systematic study of the effects of differentfreezing rates. Further studies, possibly including ultrastructural evaluation, are necessary to clarify this point. For widespread application of platelet freezing, uncontrolled freezing has obvious appeal in terms of simplicity and reduction of equipment costs. The methods described in this report, although easily performed in an investigative laboratory, may be too complex to be incorporated on a large scale into routine blood bank procedure. This is also a criticism of a rapid freezing technique (30°C/min) employing a glycerol-glucose cryoprotective medium which has recently been described (Dayian & Rowe, 1976). Future investigation should be directed towards simplifying certain of the steps, with particular attention paid to the slow addition of the DMSO-plasma and the necessity for controlled freezing rates. Comparison of transfusion results using multiple units of fresh and frozen platelets and the same donor recipient pairs is the ideal way to study such technical variations. ACKNOWLEDGMENTS

We would like to express our appreciation to Mr James Reilly, Mr Sean Hedman and Ms Theda Strohecker for their excellent technical assistance. This paper was presented in part at the Annual Meeting of the Arne rican Society of Hematology, Dallas, Texas, December 1975. REFERENCES

AISNER, J., SCHIFFER, C.A., WOLFF, J.H. & WDNIK, P.H. (1976) A standardized technique for efficient platelet and leukocyte collection using the Model 3 0 Blood Processor. Transfusion, 16, 437. ASTER, R.H. & JANDL, J.H. (1964) Platelet sequestration in man. I. Methods. Journal of Clinical Investigation, 43, 843.

BULL,B.S., SCHNEIDERMAN, M.A. & BRECEIER, G. (1965) Platelet counts with the Coulter Counter. American Journal of Clinical Pathology, 44, 678.

DAYIAN, G. & ROUTE, A.W. (1976) Cryopreservation of human platelets for transfusion. A glycerolglucose, moderate rate cooling procedure. Cryobiology, 13, I. DJERASSI, I., FARBER, S., ROY,A. & CAVWS, J. (1966) Preparation and in viuo circulation of human platelets preserved with combined dimethylsulfoxide and dextrose. Tran&sion, 6, 572. HANDIN,R.I. & VALERI,C.R. (1972) Improved viability ofpreviouslyfrozen platelets.Blood,40,509.

Frozen Platelet Transfirsion KALMBACH, K. & MARDINEY, M.R., JR (1972) An improved system for controlled rate cooling of biological material. Cryobiology, 9, 572. KIM, B.K. & BALDINI,M.G. (1973) Preservation of viable platelets by freezing. Effect of plastic containers. Proceedings of the Society .for Experimental Biology and Medicine, 142, 345. KIM, B.K. & BALDINI,M.G. (1974) Biochemistry, function, and hemostatic effectiveness of frozen human platelets. Proceedings of the Societyfor Experimental Biology and Medicine, 145, 830. KIM, B.K., TANOUE,K. & BALDINI,M.G. (1974) Short term storage at 22OC of previously frozen human platelets. Proceedings of the American Society of Hematology, p 273. MIELKE,C.H., JR, KANESHIRO, M.M., MATHER,I.A., WEINER,J.H. & RAPAPORT, S.I. (1969) The standardized Ivy bleeding time and its prolongation by aspirin. Blood, 34, 204. MURPHY,S., SAYAR,S.N., ABDOU,L. & GARDNER F.H. (1974) Platelet preservation by freezing. Use of dimethyl sulfoxide as cryoprotective agent. Trangurion, 14,139. D.H. & WIERNIK,P.H. SCHIFFER, C.A., BUCHHOLZ,

3 85

(1974) Intensive multi-unit platelet-pheresis of normal donors. Trangusion, 14, 388. SCHIFFER, C.A., WHITAKER,C.L., SCHMUKLER, M., AISNER,J. & HILBERT,S.L. (1976a) The effect of dimethyl sulfoxide (DMSO) on in vitro platelet function. Thrombosis and Huemostusis (in press). SCHIFFBR, C.A., BUCHHOLZ, D.H., AISNER, J., WOLFF J.H. & WIRNIK, P.H. (1976b)Frozen autologous platelets in the supportive care of patients with leukemia. Transfusion, 16, 321. SLICHTER, S.J. & HARKER, L.A. (1972) Cryopreservavation of viable and functional platelet concentrates. (Abstract). Clinical Research, 20, 571. SZYMANSKI, LO., PATTI,K. & KLIMAN,A. (1973) Efficacy of the Latham Blood Processor to perform plateletpheresis. Tran+ion, 13, 405. VALERI, C.R. (1974) Hemostatic effectivenessofliquidpreserved and previously frozen human platelets. New Englandlournal ofMedicine, 290, 353. VALERI,C.R., FEINGOLD, H. & MARCHIONNI, L.D. (1974) A simple method for freezing human platelets using 6% dimethylsulfoxide and storage at - 80°C. Blood, 43, 131.

Clinical experience with transfusion of cryopreserved platelets.

British J O U ~ Aof~ Haematology, 1976, 34, 377. Clinical Experience with Transfusion of Cryopreserved Platelets CHARLES A. SCHIFFER, JOSEPH AISNER...
618KB Sizes 0 Downloads 0 Views