DONOR INFECTIOUS DISEASE TESTING

Short-term deviations in temperature during storage of plasma

at -40°C do not affect its quality _3028 1541..1545

Rebecca Cardigan, Angelika Themessl, and Margaret Garwood

I
ndications for transfusion of plasma include single
BACKGROUND: Little data are available on the suit- coagulation factor deficiencies where an appropri-
ability of frozen plasma for transfusion when stored ate concentrate is not available (Factor [F]V and
outside its normal temperature, which is the focus of FXI), acute disseminated intravascular coagula-
investigation in this study. tion, plasma exchange for thrombotic thrombocytopenic
STUDY DESIGN AND METHODS: Plasma was pooled purpura, and in some instances of warfarin reversal, liver
and split to create 10 identical units on each of 24 disease, cardiopulmonary bypass, and massive transfu-
occasions (12 Group A and O). Plasma was frozen and sion.1 In the United States, fresh-frozen plasma (FFP) is
stored at -40°C for 2 weeks and then one of each of defined as plasma that is frozen within 6 to 8 hours of
the 10 identical units was subjected to one of the fol- collection; plasma frozen within 24 hours of phlebotomy
lowing deviations in storage temperature: -18°C for 1 is a separate product.2 European guidelines do not differ-
week, 1 month, or 2 months; -10°C for 1 week, 1 entiate between the two, with the term FFP being used for
month, or 2 months; 4°C for 4, 24, or 72 hours; or plasma frozen within 18 to 24 hours of collection.3 The
stored at -40°C (control) before returning all units to shelf life of frozen plasma for transfusion is determined by
-40°C. the temperature at which it is stored. European guidelines
RESULTS: Factor VIII was only significantly reduced permit FFP to be stored for up to 36 months below -25°C
when plasma was stored at 4°C for 24 hours or more and 3 months between -18 and -25°C,3 and UK guidelines,
or -10°C for 1 week. For all other arms of the study, for 24 months below -30°C.4 AABB guidelines permit FFP
the majority of units of plasma (>75%) remained above or plasma frozen within 24 hours of phlebotomy to
0.70 IU/mL and more than 95% were above the lower be stored below -18°C for up to 12 months and FFP to be
limit of normal (0.50 IU/mL). The prothrombin time ratio stored below -65°C for up to 7 years.2
only increased after storage at -10°C for 1 month or Several studies have examined the stability of plasma
more, and the activated partial thromboplastin time ratio while frozen at various temperatures at which it might
after storage at 4°C for 24 hours. None of the devia- be stored5,6 or the stability of plasma once thawed.7,8
tions in storage resulted in a decrease in fibrinogen However, there are very little data on the quality of plasma
activity. while stored frozen that is subjected to periods outside
CONCLUSION: These data suggest that plasma that its normal storage temperature. Such incidents may
has been stored at -40°C and exposed to storage tem- occur either in the blood center or in the hospital blood
peratures and times of up to 4 hours at 4°C, 1 week at bank as a result of freezer breakdowns. The lack of data
-10°C or 2 months at -18°C meets EU guidelines and makes deciding what to do with affected units in such
is suitable for transfusion.

ABBREVIATIONS: aPTT = activated partial thromboplastin

time; PT = prothrombin time.

From NHS Blood and Transplant, Brentwood; and Southampton

University Hospitals Trust, Southampton, UK.
Address reprint requests to: Rebecca Cardigan, PhD,
FRCPath, NHS Blood and Transplant, Long Road, Cambridge
CB2 2PT, UK; e-mail: rebecca.cardigan@nhsbt.nhs.uk.
Received for publication July 23, 2010; revision received
October 18, 2010, and accepted November 22, 2010.
doi: 10.1111/j.1537-2995.2010.03028.x
TRANSFUSION 2011;51:1541-1545.

Volume 51, July 2011 TRANSFUSION 1541

CARDIGAN ET AL.

circumstances difficult and may result in plasma being Coagulation analysis was performed using an analyzer
discarded unnecessarily. and reagents from the manufacturer unless otherwise
The aim of this paired study was to assess the effect stated (ACL TOP, Instrumentation Laboratory, Warrington,
of various deviations in temperature and/or time from UK). PT and aPTT were measured using kits (RecombiPlas-
storage of plasma at -40°C on its quality. We chose to test Tin and SynthAsil, respectively, both Instrumentation
levels of FVIII as a sensitive marker of these changes, Laboratory, Warrington, UK). PT and aPTT results were
fibrinogen as this should be relatively unaffected by these expressed as a ratio to the geometric mean result of pooled
changes, and the prothrombin time (PT) and activated citrated normal plasma. Fibrinogen was measured by the
partial thromboplastin time (aPTT) as screening tests that Clauss technique using the fibrinogen-C kit and calibra-
would be capable of picking up large decreases in other tion plasma. FVIII was measured by one-stage clotting
coagulation factors. assay using STA-deficient plasma and calibration plasma
(Unicalibrator, both Diagnostica Stago, Reading, UK).
Data were compared to control units stored at
MATERIALS AND METHODS
-40°C using nonparametric one-way analysis of variance
Whole blood collection and processing (ANOVA; Kruskal-Wallis test with Dunnet’s multiple com-
Whole blood (450 ⫾ 45 mL) was collected into CPD anti- parison post hoc test) using computer software (GraphPad
coagulant (63 mL) according to Guidelines for UK Trans- Prism 5.01, GraphPad Software, Inc., San Diego, CA).
fusion Services. Since levels of FVIII are lower in Group
O donors, an equal number of Group A and Group O
RESULTS
donations were selected so that this did not bias results.
Units were centrifuged (accumulated centrifugal effect Some units of plasma fractured on thawing and therefore
8.41 ¥ 107 for 12 min at 22°C, Sorvall RC 3 BP HLRb rotor data from four replicates of all arms of the study were
centrifuge, Thermo Fisher Scientific, Waltham, MA) and removed from analysis due to incomplete data, leaving 20
processed to red blood cells (RBCs), buffy coat, and replicates (11 Group A and 9 Group O). Levels of FVIII
plasma using a component extractor (Optipress II system, declined during increasing periods of storage at either 4°C
Baxter Healthcare, Deerfield, IL) and units of plasma were or -10°C, which was most pronounced for units exposed
stored overnight at 20 to 24°C, or whole blood was held at to 4°C. This was only significantly different from the
4°C for 24 hours, followed by leukoreduction (Pall WBF3, control units stored at -40°C when frozen plasma was
Pall Biomedical, Portsmouth, UK) and separation to exposed to 4°C for 24 hours or more or -10°C for 1 week or
plasma and RBCs. Three ABO- and D-matched units of more (Table 1). For all other storage deviations studied,
plasma were pooled and split to create 10 identical pedi- levels of FVIII were not significantly altered. The loss of
atric size (approx. 50-60 mL) units of plasma for each arm FVIII compared to control units was 6, 30, and 45% after
of the study. This was repeated to give 24 replicates for exposure for 4, 24, and 72 hours at 4°C, respectively, and 7,
each arm of the study (12 Group A and 12 Group O). 17, and 25% after exposure to -10°C for 1 week, 1 month,
Plasma units were rapidly frozen and stored at -40°C and 2 months, respectively. After thawing of plasma and
for 2 weeks. One of each of the 10 identical units of plasma storing units for 24 hours at 4°C, units previously exposed
was then subjected to one of the following deviations to 4°C for 24 hours or more or -10°C for 1 month or more
in storage temperature: -18°C for 1 week, 1 month, or 2 had significantly lower FVIII levels compared with control
months; -10°C for 1 week, 1 month, or 2 months; 4°C for 4, units (Table 1). The loss of FVIII compared to control units
24, or 72 hours; or -40°C (control). After the deviation in was 5, 17, and 20% after exposure for 4, 24, and 72 hours at
storage temperature, units were all returned to storage at 4°C, respectively, and 6, 11, and 18% after exposure to
-40°C for at least 1 week and units from all arms of the -10°C for 1 week, 1 month, and 2 months, respectively.
study for a given pool were then thawed and tested Exposure of frozen plasma to 4°C for 24 hours or more
together on the same occasion. or -10°C for 1 month or more resulted in FVIII levels
immediately on thawing that were less than 0.70 IU/mL in
less than 75% of units (Table 2). There was a trend toward
Laboratory analysis fewer units with greater than 0.70 IU/mL FVIII with
Units of plasma were thawed using a warm air blowing increasing periods of storage at either 4°C or -10°C.
method used routinely in hospital blood banks (SAHARA, However, frozen plasma temporarily exposed to storage
Transmed Starsted-Gruppe, Leicester, UK). Two aliquots temperatures and/or times of up to 4 hours at 4°C, 1 week
of plasma were tested immediately for PT, aPTT, and FVIII. at -10°C, or 2 months at -18°C contained more than 0.70
An additional aliquot was frozen at -70°C for analysis of IU/mL FVIII in more than 75% of units and also more than
fibrinogen at a later time. Once sampled, units of plasma 95% of these units were above the lower limit of normal for
were stored at 2 to 6°C and sampled 24 hours after thawing FVIII (0.50 IU/mL). As would be predicted, the number of
for an additional assessment of FVIII. units above either 0.70 or 0.50 IU/mL was lower in Group

1542 TRANSFUSION Volume 51, July 2011

 DEVIATIONS IN STORING FFP

TABLE 1. FVIII levels (IU/mL) in units of frozen plasma exposed to deviations in storage temperature*
Immediately on thawing 24 hr after thawing (stored at 4°C)
Deviation All units (n = 20) Group A (n = 11) Group O (n = 9) All units (n = 20) Group A (n = 11) Group O (n = 9)
Control (-40°C) 0.84 (0.76-0.92) 0.91 (0.80-1.02) 0.76 (0.63-0.88) 0.65 (0.58-0.72) 0.69 (0.59-0.79) 0.60 (0.50-0.70)
4°C 4 hr 0.79 (0.71-0.87) 0.84 (0.73-0.94)a 0.74 (0.60-0.88) 0.62 (0.55-0.69) 0.67 (0.57-0.77) 0.56 (0.45-0.68)
4°C 24 hr 0.57 (0.49-0.64)c 0.62 (0.50-0.73)c 0.51 (0.39-0.62)c 0.54 (0.47-0.62)c 0.62 (0.53-0.71)b 0.45 (0.35-0.55)c
4°C 72 hr 0.47 (0.41-0.53)c 0.52 (0.42-0.61) c
0.42 (0.34-0.50)c 0.52 (0.45-0.58) c
0.56 (0.46-0.66) c
0.46 (0.35-0.57)c
-10°C 1 week 0.76 (0.68-0.84)c 0.82 (0.72-0.93)b 0.68 (0.57-0.80)a 0.61 (0.55-0.68) 0.66 (0.57-0.75) 0.55 (0.46-0.64)
-10°C 1 month 0.68 (0.60-0.76)c 0.74 (0.63-0.84)c 0.61 (0.50-0.73)c 0.58 (0.52-0.64)c 0.61 (0.54-0.69)b 0.53 (0.43-0.64)b
-10°C 2 months 0.62 (0.55-0.69)c 0.66 (0.57-0.75)c 0.57 (0.46-0.68)c 0.53 (0.47-0.59)c 0.56 (0.49-0.64)c 0.49 (0.40-0.58)c
-18°C 1 week 0.84 (0.75-0.92) 0.92 (0.80-1.03) 0.74 (0.63-0.86) 0.66 (0.59-0.73) 0.73 (0.63-0.82) 0.58 (0.47-0.68)
-18°C 1 month 0.82 (0.73-0.91) 0.89 (0.77-1.00) 0.74 (0.60-0.87) 0.64 (0.57-0.70) 0.69 (0.60-0.77) 0.57 (0.47-0.68)
-18°C 2 months 0.82 (0.74-0.90) 0.89 (0.79-0.99) 0.73 (0.60-0.86) 0.63 (0.57-0.69) 0.68 (0.60-0.76) 0.57 (0.47-0.67)
* Data are reported as mean with 95% confidence interval. ap < 0.05, bp < 0.01, cp < 0.005 denotes p value from one-way ANOVA versus
control.

TABLE 2. FVIII levels immediately after thawing units of FFP exposed to deviations in storage temperature
compared to UK specification or lower limit of normal for FVIII (n = 20)
Percentage of units >0.70 IU/mL* Percentage of units >0.50 IU/mL
Deviation in storage from -40°C All units Group A Group O All units Group A Group O
(time and temperature) (n = 20) (n = 11) (n = 9) (n = 20) (n = 11) (n = 9)
-40°C control 80 90 66 100 100 100
4°C 4 hr 75 90 56 100 100 100
4°C 24 hr 25 36 11 65 63 56
4°C 72 hr 5 9 0 40 45 33
-10°C 1 week 75 90 56 100 100 78
-10°C 1 month 50 63 33 85 90 78
-10°C 2 months 30 36 22 80 90 66
-18°C 1 week 80 90 66 100 100 100
-18°C 1 month 80 90 66 95 100 100
-18°C 2 months 80 90 66 95 100 100
* UK specification requires greater than 75% units to contain more than 0.70 IU/mL.

O donations compared with Group A. However more than on the length of time and temperature to which frozen
75% of Group O donations were more than 0.50 IU/mL plasma was subjected during storage. Another study has
in units exposed to 4 hours at 4°C, 1 week at -10°C, or shown that levels of FVIII are reduced when plasma stored
2 months at -18°C (Table 2). at -65°C is exposed to 1 to 6°C for 2 to 4 or 24 hours by 20
The aPTT ratio was only increased after exposure to and 30%, respectively,9 higher than the 6 and 30% found in
4°C for 24 hours or more (Fig. 1B). The PT ratio was only our study. The units in the study by Dzik and colleagues9
increased when units were exposed to -10°C for 1 month were frozen within 6 hours and thus had higher starting
or longer (Fig. 1A). None of the deviations in storage values for FVIII than ours; this would make their study
from -40°C resulted in a decrease in fibrinogen activity; more sensitive to losses in FVIII during short deviations
however, measured levels were apparently higher when in temperature.
units were exposed to 4°C for 24 hours or more (Fig. 2). Current EU guidelines require that levels of FVIII in
FFP once thawed are more than 70% of the freshly collected
plasma unit.3 We did not assess plasma in this study either
DISCUSSION
at the point of whole blood donation or after the initial
A number of factors influence the level of FVIII in FFP separation of plasma from whole blood. Since 1 U of FVIII
including the anticoagulant used for collection, normal activity is usually defined as that present in 1 mL of pooled
variability between donors, and also their ABO group, the fresh normal citrated plasma, one could assume that the
time, and temperature at which whole blood or plasma is starting levels are 1 U/mL and that therefore 70% of this is
stored before freezing, the speed with which it is frozen, on average 0.70 IU/mL. In the UK more than 75% of FFP
whether any interventions such as leukoreduction or units must contain more than 0.70 IU/mL FVIII.4 Exposure
pathogen inactivation of plasma have been used during of frozen plasma to 4°C for 24 hours or more or -10°C for 1
processing, and the temperature at which it is stored. In month or more resulted in FVIII levels immediately on
this study we observed a decrease in the FVIII activity thawing that would not meet that criterion. However, those
of plasma once immediately thawed that was dependent exposed to 4 hours at 4°C, 1 week at -10°C, or 2 months at

Volume 51, July 2011 TRANSFUSION 1543

CARDIGAN ET AL.

A -18°C met UK specifications for FVIII content and also

1.5 ***
***
more than 95% of these units were above the lower limit of
normal for FVIII (0.50 IU/mL). The percentage of units that
1.0 will have a FVIII content above a specified limit depends on
PT ratio

the distribution of ABO groups, since FVIII levels are lower

in Group O donors. The plasma units in our study were 45%
0.5 Group O and 55% Group A donations, broadly reflecting
the blood group distribution of plasma that is issued.
Twenty-four hours after thawing plasma, levels of FVIII
0.0
were further reduced as would be expected, and this was
4 °C 24 hours
4 °C 72hours
4 °C 4 hours

-10 °C 1 week

-18 °C 1 week
-10 °C 2 months

-18 °C 2 months
-40 °C control

-10 °C 1 month

-18 °C 1 month
greater in units that had been exposed to 4°C for 24 hours
or more or -10°C for 1 month or more.
We also assessed levels of fibrinogen and the PT
and/or aPTT. Consistent with the loss of FVIII, the aPTT
ratio was only increased after exposure to 4°C for 24 hours
B or more. Our data are consistent with those of Dzik and
1.5
* colleagues9 who observed a 3-second prolongation of the
aPTT after exposure of 24 hours at 1 to 6°C. The PT ratio
aPTT ratio

1.0 was only increased when units were exposed to -10°C

for 1 month or longer, presumably due to the loss of
PT-dependent clotting factors FII, FV, FVII, or FX. It is not
0.5
apparent why these would be affected by the deviation at
-10°C and not at 4°C. Others have shown that exposure of
0.0 plasma stored at -80°C to 4°C for 24 hours on two occa-
4 °C 24 hours
4 °C 72hours
-10 °C 1 week

-18 °C 1 week
4 °C 4 hours
-40 °C control

-10 °C 2 months

-18 °C 2 months
-10 °C 1 month

-18 °C 1 month

sions results in less than a 10% decrease in the vitamin

K–dependent factors FII, FVII, F IX, and FX.10 None of the
deviations in storage from -40°C in our study resulted in a
decrease in fibrinogen activity as might be expected, since
this protein is known to be relatively stable in plasma.
Fig. 1. aPTT (A) and PT (B) ratios in units of frozen plasma However, measured levels were apparently higher when
exposed to deviations in storage temperature. Data are units were exposed to 4°C for 24 hours or more, which
presented as the median with range (n = 20). *p < 0.05 could be caused by cold activation of some coagulation
and ***p < 0.001 versus control unit stored at -40°C. factors in plasma.
Some elements in the design of our study mean that
6
the absolute levels of FVIII in the units are worst case. First,
*** *
we used plasma that was prepared after separation from
Fibrinogen (g/L)

whole blood on the day of collection and then stored for 1

4 day at ambient temperature or separated from whole blood
after storage at 4°C, neither of which are currently standard
practice in England. This was to avoid wasting units
2 of plasma destined for transfusion. As a result, the absolute
starting levels of FVIII observed in units of plasma in this
study are lower than those in routine production (usually
0 0.90-1.00 IU/mL on average in England). Second, to use
4 °C 24 hours
4 °C 72hours
4 °C 4 hours

-10 °C 1 week

-18 °C 1 week
-10 °C 2 months

-18 °C 2 months
-40 °C control

-10 °C 1 month

-18 °C 1 month

identical units in all arms of the study, we used pediatric-

sized units of plasma, which are approximately 50 to 60 mL
in volume compared with an adult size unit of approxi-
mately 300 mL. It is likely that the effect of the deviations in
time and temperature that we observed would be less
Fig. 2. Fibrinogen levels in units of FFP exposed to deviations marked in adult units, since these would take longer to
in storage temperature. Box and whisker plots represent the reach a higher temperature. Therefore, our conclusions
median and 25th and 75th percentiles and minimum and based on our data are probably conservative.
maximum values (n = 20). *p < 0.05 and ***p < 0.001 versus Given the current indications for the use of plasma for
control unit stored at -40°C. transfusion, do the decreases observed in our study
matter clinically? There is a paucity of data demonstrating

1544 TRANSFUSION Volume 51, July 2011

 DEVIATIONS IN STORING FFP

the efficacy of plasma for many of the clinical uses CONFLICT OF INTEREST
for which it is indicated, with the exception of plasma
None of the authors disclose any conflicts of interest.
exchange for thrombotic thrombocytopenic purpura.11 It
is therefore difficult to determine how the results of in
REFERENCES
vitro analysis of plasma relate to its clinical efficacy. We did
not measure all the coagulation factors in plasma that are 1. O’Shaughnessy DF, Atterbury C, Bolton Maggs P, Murphy
probably important for maintaining hemostasis, nor their M, Thomas D, Yates S, Williamson LM; British Committee
inhibitor proteins since FVIII would be expected to be for Standards in Haematology, Blood Transfusion Task
the most affected factor. This is presumably why routine Force. Guidelines for the use of fresh-frozen plasma,
quality monitoring of FFP in Europe requires FVIII to be cryoprecipitate and cryosupernatant. Br J Haematol 2004;
measured on an ongoing basis, since it is the most sensi- 126:11-28.
tive marker of alterations in the storage time and/or tem- 2. AABB Standards Program Committee. Standards for blood
perature of blood before plasma production and also of banks and transfusion services. 26th ed. Bethesda (MD):
plasma once frozen. Further, measurement of FVIII is a American Association of Blood Banks; 2009.
routine assay in hematology laboratories for which inter- 3. Council of Europe. Guide to the preparation, use and
national standards exist. However, if the time and/or tem- quality assurance of blood components. 14th ed.
perature of plasma production is well controlled it is not Strasbourg: Council of Europe Publishing; 2009.
apparent what extra assurance for patients measuring 4. James V, editor. Guidelines for the blood transfusion
FVIII routinely in FFP provides. In our study, frozen services in the United Kingdom. 7th ed. London: The
plasma that has been stored at -40°C and temporarily Stationery Office; 2005.
exposed to storage temperatures and/or times of up to 4 5. Koerner K, Stampe D. [Stability of blood coagulation
hours at 4°C, 1 week at -10°C, or 2 months at -18°C con- factors in deep frozen fresh plasma by storage at -20
tained normal levels of FVIII in more than 95% of units, degrees C and -40 degrees C]. Infusionsther Klin Ernahr
still meets current EU/UK specifications for FVIII content 1984;11:46-50.
on thawing, and had unaltered PT and/or aPTT or fibrino- 6. Riggert J, Morsdorf S, Pindur G, Köhler M, Seyfert UT.
gen. Based on these findings we suggest that plasma Quality of fresh-frozen plasma during storage. Vox Sang
exposed to such deviations may be suitable for transfu- 1997;73:257.
sion. It is generally accepted that a level of FVIII of 0.30 to 7. Downes KA, Wilson E, Yomtovian R, Sarode R. Serial mea-
0.40 IU/mL is required to maintain normal hemostasis. surement of clotting factors in thawed plasma stored for 5
FFP is no longer used to correct FVIII deficiency in days. Transfusion 2001;41:570.
patients with hemophilia, but mainly transfused in an 8. Yazer MH, Cortese-Hassett A, Triulzi DJ. Coagulation factor
attempt to correct multiple coagulation factor deficien- levels in plasma frozen within 24 hours of phlebotomy over
cies. Most patients for whom FFP is indicated in fact have 5 days of storage at 1 to 6 degrees C. Transfusion 2008;48:
normal or high FVIII levels and it would therefore seem 2525-30.
reasonable to suggest that FVIII levels in excess of 0.3 9. Dzik WH, Riibner MA, Linehan SK. Refreezing previously
IU/mL in FFP might be adequate. In our study, even thawed fresh-frozen plasma. Stability of coagulation
frozen plasma exposed to the worst temperature and/or factors V and VIII:C. Transfusion 1989;29:600-4.
time deviation and then thawed and stored for an addi- 10. Ben Tal O, Zwang E, Eichel R, Badalbev T, Hareuveni M.
tional 24 hours contain more than 0.30 IU/mL FVIII in Vitamin K-dependent coagulation factors and fibrinogen
95% of units. It is possible that plasma exposed to such levels in FFP remain stable upon repeated freezing and
temperatures may be clinically acceptable despite not thawing. Transfusion 2003;43:873-7.
meeting current EU standards. 11. Stanworth SJ, Brunskill SJ, Hyde CJ, McClelland DB,
We did not assess plasma that has been pathogen Murphy MF. Is fresh frozen plasma clinically effective?
inactivated using one of the three licensed systems A systematic review of randomized controlled trials.
available in Europe for single-unit plasma: riboflavin, Br J Haematol 2004;126:139-52.
amotosalen, or methylene blue. Given that these systems 12. Hornsey VS, Drummond O, Morrison A, McMillan L,
themselves result in a reduction in the level of FVIII in MacGregor IR, Prowse CV. Pathogen reduction of fresh
FFP,12,13 one must bear in mind that the conclusions from plasma using riboflavin and ultraviolet light: effects on
our study may not be true for treated plasma. plasma coagulation proteins. Transfusion 2009;49:2167-72.
13. Osselaer JC, Debry C, Goffaux M, Pineau J, Calomme G,
Dubuc E, Chatelain B, Vandendaele MC, Hsu J, Rhein-
ACKNOWLEDGMENTS
schmidt M, Lin L. Coagulation function in fresh-frozen
This study formed part of an MSc project in Transfusion and plasma prepared with two photochemical treatment
Transplantation Science at the University of Bristol, Bristol, methods: methylene blue and amotosalen. Transfusion
England. 2008;48:108-17.

Volume 51, July 2011 TRANSFUSION 1545