BLOOD COMPONENTS
Noninvasive pH monitoring of platelet concentrates:
a large field test
Effimia Gkoumassi, Christa Klein-Bosgoed, Margriet J. Dijkstra-Tiekstra, Dirk de Korte, and
Janny de Wildt-Eggen
BACKGROUND: Developing new quality control
methods for platelet concentrates (PCs) can contribute
to increasing transfusion safety and efficiency. The aim
of this study was to investigate in a large field test the
quality of expired PCs and whether 100% noninvasive
pH monitoring can be used to predict PC quality.
STUDY DESIGN AND METHODS: The pH of 13,693
PCs produced for transfusion was monitored daily using
Blood Storage, Inc.’s pH sterile, automated fluoroscopic
evaluation technology. Upon indication of compromised
quality or expiration, PCs were returned and in vitro
tests were performed.
RESULTS: A total of 998 PCs were returned, of which
962 outdated, 26 had a positive BacT/ALERT reaction,
seven had aggregates, one was without swirl, one had
low pH, and one had high pH. BacT/ALERT was faster
in identifying bacterial contamination than pH measurements. The pH at the end of the storage period was
significantly lower than at the beginning. In vitro tests
indicated that while the PC quality was acceptable upon
expiration, it rapidly declined after expiration.
CONCLUSION: In this setting where the vast majority
of PCs were of good quality and within acceptable pH
limits, daily, noninvasive routine pH measurement has
limited added value in identifying quality-compromised
PCs.
P
latelets (PLTs) are sensitive to pH variations and
the ideal storage pH37°C is between 6.8 and 7.4,
although acceptable storage limits are set as
pH37°C > 6.3 (pH22°C 6.4).1,2 A pH below 6.4 is associated with a lower PLT viability.3 Rapid lactate accumulation can occur in a PLT concentrate (PC) with high PLT
content and consequently pH can quickly decrease.4
A rapid decrease in pH can also be the consequence
of bacterial contamination with acid-producing bacteria.5
Bacterial contamination of blood products can cause
bacterial-induced transfusion-related sepsis, which in
turn increases patient transfusion-related morbidity and
mortality in patients. Despite recently introduced safety
measures, bacterial contamination still represents one of
the most frequent infectious risks of transfusion.6-8
Storage lesion of PLTs can also be induced through
poor storage conditions, such as gas-impermeable storage
bags, high or low temperatures, too rigorous shaking, or
absence of shaking for more than 24 hours.3,9-13 Consequently, monitoring the pH and assuring that it is within
acceptable limits could be of added value to blood banking.
Most methods of pH measurement (blood gas analyzers
[BGA], handheld pH meters, and colorimetric dipsticks)
are invasive, expensive, laborious, and inaccurate (as in
the case of dipsticks) or require a large sample volume,
ABBREVIATIONS: BCSI = Blood Storage, Inc.; BGA(s) = blood
gas analyzer(s); PC(s) = platelet concentrate(s); SAFE = sterile,
automated fluoroscopic evaluation.
From the Department Transfusion Monitoring, Sanquin
Research, Groningen; and the Department Product and Process
Development, Sanquin Blood Supply, Amsterdam, the
Netherlands.
Address reprint requests to: Effimia Gkoumassi, Department
Transfusion Monitoring, Sanquin Research, PO Box 1191,
9701 BD Groningen, the Netherlands; e-mail:
E.Gkoumassi@gmail.com.
Received for publication July 25, 2012; revision received
November 29, 2012, and accepted November 30, 2012.
doi: 10.1111/trf.12099
TRANSFUSION 2013;53:2287-2292.
Volume 53, October 2013 TRANSFUSION
2287
GKOUMASSI ET AL.
resulting in product wastage. A new system, Blood Storage,
Inc. (BCSI)’s pH sterile, automated fluoroscopic evaluation
(SAFE) reader (also known as the pH 1000; BCSI, Seattle,
WA), offers the ability to monitor pH in a noninvasive
manner, which could be used on 100% of PCs to identify
those of compromised quality.5,7,14,15
Over a period of 15 months (June 2009 to September
2010), a large field test was performed with PCs stored in a
bag containing an integrated pH sensor, allowing noninvasive pH monitoring through the use of the BCSI pH
SAFE reader until just before transfusion. Furthermore,
the quality of PCs after expiration was investigated as well
as whether pH could be used to identify PCs unsuitable for
transfusion.
MATERIALS AND METHODS
PC preparation
Whole blood–derived, multiple-donor PCs were made
for clinical use, complying with European guidelines,
using validated production protocols, as previously
described.2,16 In short, PCs were made by pooling the buffy
coats of five volunteer donors together with 1 plasma unit
from one of those donors. Subsequently this pool was
gently centrifuged (900rcf; ACE 1.27 ¥ 109, Break 3) in a
centrifuge (Sorvall 12BP, Thermo Fisher Scientific, Inc.,
Waltham, MA) and the resulting PLT-rich plasma was
slowly extracted with an automated separator (Compomat
G4, Fresenius Kabi, Emmer-Compascuum, the Netherlands) into a polyvinylchloride-citrate storage bag with an
integrated BCSI pH probe. The PC was stored in a shaking
incubator (one cycle per second) at 22 ⫾ 2°C until transfusion (maximally 7 days after production) or in vitro
testing (in case of expiration or recall; maximally 11 days
after production). There was no quarantine time for the
PCs before release.
Quality control
A bacterial detection system (BacT/ALERT 3D system,
bioMérieux, Marcy l’Etoile, France) was used for bacterial
contamination screening of all PCs, with incubation of
7.5 mL of concentrate in an aerobic and 7.5 mL in an
anaerobic bottle for 7 days at 35°C.
PLT concentration was measured using a hematology
analyzer (Sysmex XT1800i, Toa, Japan) and used to calculate the amount of PLTs per concentrate in combination
with the volume. Only PCs with more than 250 ¥ 109 PLTs,
the presence of swirling effect, and negative-to-date for
bacterial contamination were released for transfusion.
From Day 2 onward and until transfusion, the pH of
the PLTs was measured daily (if possible) using the BCSI
pH SAFE reader. The measurements took place at the
blood bank, three blood bank distribution centers, and
four hospitals (Isala Klinieken, Zwolle; Medisch Centrum
2288
TRANSFUSION Volume 53, October 2013
Leeuwarden; Medisch Centrum Twente, Enschede; Universitair Medisch Centrum, Groningen, the Netherlands).
PCs, either in the blood bank or in the hospital inventory, were sent to the research department if they exhibited absence of swirl, presence of aggregates, pH outside
the acceptable range, or upon their expiration (7 days after
blood collection). During the study, the acceptable pH
range was pH37°C 6.3 to 7.3. Successfully recalled PCs with
a positive BacT/ALERT were also sent to the research
department for in vitro testing.
In vitro testing
All assays were performed with validated standard
methods available in our laboratory. Of all PCs returned,
the pH37°C was measured daily, on the day of return and on
Day 9, 10, or 11, using the BCSI pH SAFE reader, as previously described.5,14,15
In short, the BCSI pH SAFE method uses a
pH-sensitive fluorescent membrane fixed to a clear
window inside a small sensor tube which is welded in the
rim of the PC container. Fluorescence of the sensor membrane is measured through the window using a fiber optic
probe and the BCSI pH SAFE reader calculates pH based
on the ratio of yellow and red fluorescence measured.
Returned PCs were sampled using a syringe and luerlock adapter and tested on a BGA (RapidLab, Siemens
Medical Solutions Diagnostics BV, Breda, the Netherlands)
for pH37°C, pO2, pCO2, glucose, and lactate concentration
on Day 9, 10, or 11. The returned PCs were also tested for
annexin V binding and CD62P surface expression using
flow cytometry (FACSCalibur, Becton Dickinson, San Jose,
CA), as previously described.17 In case of a successful recall
due to an initial positive BacT/ALERT reaction, PCs were
first aseptically sampled and then cultured for a second
time in BacT/ALERT, before proceeding to in vitro testing.
Only the PCs where the second sample gave a positive
BacT/ALERT reaction were considered to be confirmed
positively contaminated.
Statistical analysis
Results are presented as mean ⫾ standard deviation (SD).
Significance was determined using unpaired t test, and
analysis of variance (with Tukey posttest) with a p value of
less than 0.05 used to indicate significance. All statistical
tests were performed using commercially available software (Microsoft Excel, Microsoft Corp., Redmond, WA;
Instat, GraphPad Software, La Jolla, CA).
RESULTS
Produced, returned, and recalled PCs
In total, 13,693 PCs were produced during the period of
the study. The distribution of PCs per study hospital was
NONINVASIVE pH MONITORING OF PCs
as follows: 20% to Isala Klinieken, Zwolle; 7% to Medisch
Centrum, Leeuwarden; 30% to Medisch Centrum Twente,
Enschede; and 43% to Universitair Medisch Centrum,
Groningen. On the production day (Day 1) PLT count
(341 ¥ 109 ⫾ 53 ¥ 109) and volume (346 ⫾ 17 mL) were in
compliance with national guidelines. During the storage
and distribution of the PCs (Day 2 to 7), there were
37,307 pH readings collected with the pH SAFE readers.
Of all PCs, 998 were returned: 962 due to outdating, 26
recalled successfully due to an initial positive BacT/
ALERT reaction, seven due to presence of aggregates (pH
was >7.0 and swirl normal in all seven cases), one due to
absence of swirl, one due to low pH (<6.2), and one due
to high pH (pH 7.42). No adverse transfusion reactions to
PC transfusions or cases of transfusion-related sepsis
were reported over the course of the study.
Low and high pH37°C measurements
Low pH (pH37°C < 6.3) was measured on the BCSI pH SAFE
reader in only 10 (0.073%) of all PCs before expiration
(Days 2-7) and four of these were also confirmed with
positive bacterial contamination (0.029%). One of the PCs
(0.007%) that developed a low pH37°C (pH37°C 6.58 on Day 3,
return day; pH37°C < 6.3 on Day 6) contained a high
amount of PLTs (629 ¥ 109 per bag) and was returned due
to absence of swirl. One PC with a characteristic deviating
brown color was returned due to low pH (on Day 7; BCSI
pH SAFE value < 6.3). The following day the pH was determined on the BGA (pH 7.133), indicating that it still fell
within the acceptable range, although not according to the
BCSI pH SAFE system (pH < 6.3). No positive culture reaction was detected, swirl was normal, and all in vitro tests
(performed in this case on Day 8) gave normal results.
Consequently, it was speculated that the brown color
interfered with the BCSI pH SAFE measurement.
For two PCs (0.015%) low pH37°C measurements could
be explained by handling error of the operator or
mechanical fault of the apparatus, based on the fact that a
repeated pH measurement gave normal results. Two cases
were unexplainable. PLT content was not deviating and
since the PCs were transfused no further in vitro tests were
performed. No adverse transfusion reactions were
reported. All five PCs with confirmed low pH (four with
confirmed bacterial contaminations and one with high
PLT content) were of extremely compromised quality and
unsuitable for transfusion.
Over the whole study, one PC was returned as a result
of a high pH. On Day 4, BCSI pH SAFE indicated a pH value
of 7.42, which was above the desirable upper limit. The
swirl and in vitro tests (performed on Day 9) were normal
and no interfering variables could be identified for this PC.
after an initial increase (on Day 3 pH was 7.31 ⫾ 0.09).
However, the decrease was only minimal from Day 4 to
Day 7 (Fig. 1A). The main decrease was found after expiration, between Day 7 and Day 10 (Fig. 1A) but pH
remained greater than 6.8 in most cases. On Day 11 the
pH was significantly higher than on Day 10, probably
because not all PCs were measured on all days.
pH37°C BCSI pH SAFE versus pH37°C BGA
Upon returning to the laboratory, the in vitro tests were
performed and the pH37°C was measured on the same day
using both the BCSI pH SAFE reader and the BGA. The pH
measured with the BCSI pH SAFE reader (7.09 ⫾ 0.009)
was significantly higher than the pH measured with the
BGA (6.99 ⫾ 0.006; p < 0.05), implying that the BCSI pH
SAFE readers used in this test slightly overestimated the
pH.
In vitro tests of expired PCs
In vitro tests of expired PCs were performed in most cases
on Day 9, 10, or 11. On Day 9 (n = 540) the mean CD62P
expression and annexin V binding were significantly
lower than on Day 10 (n = 79) and Day 11 (n = 250), but
no significant difference was found between Day 10 and
Day 11 (Fig. 2A). CD62P expression was greater than 50%
for 15 (2.8% of PCs), 18 (23.4% of PCs), and 23 (9.7% of
PCs) PCs on Days 9, 10, and 11, respectively. Annexin V
binding was greater than 40% for 44 (8.2% of PCs), 24
(31.2% of PCs), and 69 (29.2% of PCs) PCs on Days 9, 10,
and 11, respectively.
In accordance with reduced but still active metabolism, pO2 was higher on Day 11 than on Day 9, pCO2 was
significantly lower on Day 11 than on Days 9 and 10.
Glucose was significantly higher and lactate concentration was significantly lower on Day 9 than on Days 10 and
11 (Fig. 2B). Lactate was greater than 20 mmol/L in 23% of
the PCs on Day 9, 46% of the PCs on Day 10, and 54% of the
PCs on Day 11. Glucose remained present (>5 mmol/L)
during storage for all tested PCs.
On Day 9, on average, the PCs exhibited swirl and
acceptable pH; however, CD62P expression, annexin V
binding, and lactate concentration were high. In total, at
least 125 PCs were of compromised quality on Day 9
(based on in vitro tests). If Day 7 pH was less than 7.1,
CD62P expression, annexin V binding, and lactate concentration on Day 9 were higher than for units with
pH values of greater than 7.1 (n = 12). However, due to the
low sample size, it cannot be claimed that pH on Day 7 is
predictive of the PC quality on Day 9.
pH progress during storage
BacT/ALERT and pH
In accordance with previous studies, pH decreased during
storage from pH 7.27 ⫾ 0.08 on Day 2 to pH 7.08 ⫾ 0.21
Of all PCs, 48 (0.351%) gave an initial positive culture reaction of which 26 were successfully recalled. Of these 26
Volume 53, October 2013 TRANSFUSION
2289
GKOUMASSI ET AL.
A
7.4
7.2
pH37°C
7.0
6.8
6.6
6.4
6.2
2
9993
3
8964
4
6871
5
5435
6
4060
7
1984
8
795
9
598
10
96
11
567
Day
Units tested per day (n)
B
7.4
pH37°C
7.2
*
*
*
**
7.0
*
6.8
6.6
6.4
6.2
2
3
4
5
6
7
8
9
Day
Fig. 1. (A) The pH37°C during storage in “normal” PCs (mean ⫾ SD). Not all PCs were measured all days. X-axis: top row, PC age in
days; bottom row, number per day. (B) Mean pH decrease during storage in normal PCs (䉱; number as in A), PCs that gave a positive BacT/ALERT reaction (•; n = 48) and PCs that were confirmed positive (䊏; n = 9).*p < 0.05 normal PCs versus BacT/ALERT positive PCs.
PCs, nine (0.066%) were confirmed positive by a second
culture, with all confirmed positive cases originating from
skin flora microorganisms. The pH remained within the
acceptable range on the day of the first positive culture
reaction; however, in vitro tests indicated deviating values
(data not shown).
Three PCs gave a positive reaction within 24 hours of
inoculation (Day 2) and one of these also exhibited
pH37°C of less than 6.3 (Staphylococcus aureus). In total,
the pH of four bacterially infected PCs (Bacillus
sp., S. aureus, S. aureus, and hemolytic Streptococcus)
dropped to less than 6.3 on Days 7, 7, 2, and 4, respectively (not shown).
BacT/ALERT was approximately 3 days faster in identifying suspected contamination than pH measurements,
although it must be kept in mind that BacT/ALERT is cul2290
TRANSFUSION Volume 53, October 2013
turing a sample at elevated temperatures whereas pH is
measured in the real product to be transfused. It was
observed that pH decrease in PCs with a positive BacT/
ALERT reaction, whether confirmed positive or not, was
significantly more rapid than in the normal PCs (Fig. 1B).
The pH of confirmed positive bacterially contaminated
PCs appears to decline drastically but due to the low frequency of confirmed contaminations, the difference is not
significant (Fig. 1B).
DISCUSSION
During this large field study for monitoring pH of PCs the
great majority of all PCs were of high quality, meeting
European standards. Noninvasive pH measurement could
help with monitoring the quality of PCs containing high
initial amounts of PLTs, taking into account that a high
NONINVASIVE pH MONITORING OF PCs
A new software upgrade available
for the BCSI pH SAFE reader, developed
50
after the end of this study (and partly as a
*
result of this study), makes user calibra*
40
tion to a reference (e.g., BGA) possible
and could potentially increase accuracy.
30
It has been suggested recently
20
that present PC quality control (QC)
methods, including bacterial contami10
nation screening are insufficient, often
resulting in product wastage and conse0
quently new techniques should be
D9 D10 D11
D9 D10 D11
employed to further reduce bacterial
B
risk in PCs.20
Bacterial screening systems cannot
25
*
be substituted with pH measurements,
however, and while on the one hand it is
20
*
of interest that pH trending of potentially contaminated PCs was found to be
†
15
distinctly different than normal PCs, on
the other hand, very low concentrations
10
of microorganisms in PCs could poten‡
†
tially grow in BacT/ALERT (at 35°C) and
5
result in an initial positive reaction but
fail to survive in the PC which is stored
at a lower temperature.
0
D9 D10 D11
D9 D10 D11
D9 D10 D11
D9 D10 D11
The BCSI pH SAFE system could
be deemed suitable for use until just
Fig. 2. In vitro tests of expired PCs (mean ⫾ SD). (A) Percentage CD62P membrane
before transfusion by hospital staff, but
expression (䉬) and annexin V (䊏) binding. (B) pO2 in kPa (䉬), pCO2 in kPa (䊐),
not as a blood bank monitoring instruglucose in mmol/L (䉱), and lactate content in mmol/L (䊊) during Days 9, 10, and 11.
ment. However, it must be kept in mind
*p < 0.05 Day 9 versus Days 10 and 11. †p < 0.05 Day 9 versus Day 11. ‡p < 0.05 Day
that the overall process of whole blood
10 versus Day 9 and Day 11. Not all PCs were measured all days.
collection and processing of PCs was
very standardized in this study, leading to restricted
PLT concentration can lead to quicker depletion of
variation in the volume and PLT content of the PCs. This
glucose and a rapid increase of lactate, resulting in lower
was reflected in the results of the study, with only a
pH and reduced PLT viability. The amount of PCs prominimal amount of aberrant pH values in the PCs during
duced with high PLT concentrations was limited in the
storage for up to 7 days. The use of the pH SAFE techcurrent study, but earlier studies have shown this correlanology confirmed that the routine production and distrition.4 However, rapid deterioration of PCs containing high
bution setting of this study is well controlled and capable
amounts of PLTs can also be prevented by releasing (and
of producing PCs of high quality, with very few PCs with
transfusing) such PCs more quickly, but that requires
low pH values during their 7-day storage. Under these
100% counting of PCs.
conditions, daily pH monitoring does not have added
All in vitro variables measured on Days 9, 10, and 11
value to currently implemented QC tools (such as swirl
were in line with the reducing quality of the PCs and comand BacT/ALERT) in identifying PCs unsuitable for transparable with what is previously reported for PLTs of this
fusion. Under less standardized conditions, daily pH
age.17
monitoring might identify more deviating units and be of
The BCSI pH SAFE system was user-friendly and did
additional value.
not result in product wastage when compared to BGAs.
However, pH measurements with BGAs were more accurate and were comparable to previous studies.
ACKNOWLEDGMENTS
An AABB Standard characterizes pH, glucose concentration measurement, and microscopy as less sensitive
The authors thank the employees of the transfusion monitoring,
methods for detecting bacterial contamination of PCs and
blood processing and distribution, and client service departconsequently discourages their use, which is confirmed by
ments of Sanquin Blood Supply in Groningen for their collaboraour results.18,19
tion and E.J. Klip and S. Geelhood from BCSI for their advice.
pO2 (kPa), pCO2 (kPa), glucose
(mmol/L) and lactate (mmol/L)
% Annexin V binding
% CD62P expression
A
Volume 53, October 2013 TRANSFUSION
2291
GKOUMASSI ET AL.
CONFLICT OF INTEREST
At the time of this work DdK was a member of the scientific
advisory committee of Blood Cell Storage, Inc., Seattle, WA (BCSI).
[Correction added after online publication 30-Jan-2013. The
company name has updated.] This committee is presently inactive. The remaining authors declare that they have no conflicts of
interest relevant to the manuscript submitted to TRANSFUSION.
9. Murphy S, Gardner FH. Effect of storage temperature on
maintenance of platelet viability—deleterious effect of
refrigerated storage. N Engl J Med 1969;280:1094-8.
10. Murphy S, Sayar SN, Gardner FH. Storage of platelet concentrates at 22 degrees C. Blood 1970;35:549-57.
11. Slichter SJ, Harker LA. Preparation and storage of platelet
concentrates. II. Storage variables influencing platelet
viability and function. Br J Haematol 1976;34:403-19.
12. Tynngard N. Preparation, storage and quality control of
REFERENCES
1. Dekkers DW, De Cuyper I, van der Meer PF, Verhoeven AJ,
de Korte D. Influence of pH on stored human platelets.
Transfusion 2007;47:1889-95.
2. Council of Europe. Guide to the preparation, use and
platelet concentrates. Transfus Apher Sci 2009;41:
97-104.
13. van der Meer PF, de Korte D. Platelet preservation:
agitation and containers. Transfus Apher Sci 2011;44:297304.
14. Krause M, Doescher A, Zimmermann B, Muller TH. Nonin-
quality assurance of blood components. Reccomendation
vasive pH measurement to monitor changes during subop-
R (95) 15. 16th ed. Strasbourg: Council of Europe
Publishing; 2010.
timal storage of platelet concentrates. Transfusion 2010;50:
2185-92.
3. Murphy S, Gardner FH. Platelet storage at 22 degrees C:
15. Reed MW, Geelhood S, Barker LM, Pfalzgraf R, Vlaar R,
role of gas transport across plastic containers in maintenance of viability. Blood 1975;46:209-18.
4. Dijkstra-Tiekstra MJ, Kuipers W, Setroikromo AC, de Wildt-
Gouwerok E, De Cuyper I, Harris P, Verhoeven AJ, de Korte
D. Noninvasive measurement of pH in platelet concentrates with a fiber optic fluorescence detector. Transfusion
Eggen J. Platelet capacity of various platelet pooling
systems for buffy coat-derived platelet concentrates. Transfusion 2008;48:2114-21.
2009;49:1233-41.
16. Dijkstra-Tiekstra MJ, Kuipers W, Setroikromo AC, de WildtEggen J. Overnight or fresh buffy coat-derived platelet con-
5. Barker LM, Nanassy OZ, Reed MW, Geelhood SJ, Pfalzgraf
RD, Cangelosi GA, de Korte D. Multiple pH measurement
during storage may detect bacterially contaminated platelet concentrates. Transfusion 2010;50:2731-7.
6. de Korte D, Curvers J, de Kort WL, Hoekstra T, van der Poel
CL, Beckers EA, Marcelis JH. Effects of skin disinfection
method, deviation bag, and bacterial screening on clinical
safety of platelet transfusions in the Netherlands. Transfusion 2006;46:476-85.
7. de Korte D. 10 years experience with bacterial screening of
platelet concentrates in the Netherlands. Transfus Med
Hemother 2011;38:251-4.
8. Hillyer CD, Josephson CD, Blajchman MA, Vostal JG,
Epstein JS, Goodman JL. Bacterial contamination of blood
components: risks, strategies, and regulation: joint ASH
and AABB educational session in transfusion medicine.
Hematology Am Soc Hematol Educ Program 2003;575-89.
2292
TRANSFUSION Volume 53, October 2013
centrates prepared with various platelet pooling systems.
Transfusion 2008;48:723-30.
17. Dijkstra-Tiekstra MJ, Pietersz RN, Huijgens PC. Correlation
between the extent of platelet activation in platelet concentrates and in vitro and in vivo parameters. Vox Sang
2004;87:257-63.
18. American Association of Blood Banks. Association Bulletin
#10-05—suggested options for transfusion services and
blood collectors to facilitate implementation of BB/TS
interim standard 5.1.5.1.1. Bethesda (MD): American Association of Blood Banks; 2010.
19. American Association of Blood Banks. Association Bulletin
#10-02—interim standard 5.1.5.1.1. Bethesda (MD): American Association of Blood Banks; 2010.
20. Tomasulo P, Su L. Is it time for new initiatives in the blood
center and/or the hospital to reduce bacterial risk of platelets? Transfusion 2011;51:2527-33.