Pre-instrumental Variables in Coagulation
Testing
J O H N A. K O E P K E , M.D., J O H N L. R O D G E R S ,
AND M A R T H A
M.D.,
J. O L L I V I E R , MT (ASCP)
Department of Pathology, University of Iowa Hospitals and Clinics Iowa City, Iowa 52240
ABSTRACT
Koepke, John A., Rodgers, John L., and Ollivier, Martha J.: Pre-instrumental variables in coagulation testing. Am J Clin Pathol 64: 591-596,
1975. A number of variables thought to affect measurement of prothrombin
time (PT) and partial thromboplastin time (PTT) were examined in an effort to determine more precisely their effects on these measurements.
On the basis of these studies, it is proposed that blood specimens be anticoagulated with one part 3.8% (w./v.) sodium citrate solution to 19 parts
whole blood to avoid excessive anticoagulation of blood samples drawn from
patients with polycythemia. Because of the smaller amounts of plasma in
such samples, relatively larger amounts of anticoagulant are used, and
spuriously prolonged PT and P T T measurements commonly result. No
deleterious effect on anemic specimens is evident when the smaller amount
of citrate is used. Studies of the stability of these specimens indicate that
unopened, vacuum-drawn specimens do not noticeably deteriorate for as
long as 6 hours, even when kept at room temperature. Prothrombin time
measurements remain constant for as long as 24 hours. However, a 1 0 - 1 5 %
lengthening of the partial thromboplastin time is evident after 24 hours of
storage. (Key words: Partial thromboplastin time (PTT); Prothrombin time
(PT); Anticoagulant citrate; Hematocrit; Stability of coagulation factors.)
S I G N I F I C A N T VARIABILITY of
the two most
widely used coagulation tests, the prothrombin time and partial thromboplastin
time tests, has been reported in the CAP
surveys. 2,3 T h e reasons for such variability
are not completely identified; therefore,
this information was sought in a survey of
participants in the College of American
Pathologists Hematology Survey. This
study, done in the fall of 1973, was made to
investigate in more detail some variables
known to affect coagulation tests. In addi-
tion to documenting the wide variety of
technics used for these measurements, the
survey disclosed significant deviations from
the manufacturer's instructions in performances of the tests by many laboratory
workers. For example, 40% of the users
of the several manufacturers' partial
thromboplastin reagents failed to incubate
the reaction mixtures for the time recommended by the manufacturers. Likewise, a calcium chloride concentration
different from that recommended was
used by 12% of the responding labora-
Received January 2, 1975; received revised manuscript April 7, 1975; accepted for publication
tories.3 O t h e r e x a m p l e s Could b e given, but
,i
•
» « . • . .
»u ^
• *
v
y
April 7, 1975.
Address reprint requests to Dr. Koepke.
the important point was that appropriate
methodology is not always used, and such
591
592
KOEPKE, RODGERS, AND OLLIVIER
A.J.C.P.—Vol. 64
Table 1. Effects of Hematocrit on Prothrombin Time and Partial Thromboplastin Time
(Effects of Increasing vs. Uniform Citrate Concentration)
Increasing Citrate Concentration
(Standard Collection Procedure)
Measured
Hematocrit
Uniform Citrate Concentration
(%)
Ca + +
(mg. per dl.)
PT
(Sec.)
PTT
(Sec.)
Ca + +
(mg. per dl.)
PT
(Sec.)
PTT
(Sec.)
5
10
14
19
24
29
35
40
46
50
54
60
65
69
10.7
10.6
10.7
10.8
10.7
11.4
11.5
10.6
10.1
9.9
11.3
10.1
9.4
10.2
13.5
13.7
14.0
14.0
14.0
13.7
14.0
14.0
14.2
15.0
14.7
16.1
17.0
22.1
33.7
32.0
34.0
34.5
35.2
36.5
38.7
38.2
40.6
41.5
48.2
56.9
63.8
84.7
10.6
10.0
9.8
9.65
10.0
10.3
9.8
9.8
9.8
8.15
9.8
9.4
12.2
8.75
13.5
13.5
14.0
13.5
14.0
14.0
13.7
14.0
14.0
14.7
14.0
14.0
13.7
13.7
32.5
31.8
31.3
33.5
34.2
35.0
34.0
33.5
33.2
38.9
36.8
36.4
37.0
35.0
deviations probably significantly affect the acceptable. Because of the widespread use
results these laboratories submit in the of this concentration, coupled with the
CAP survey program.
lack of any appreciable effect on the preserFor many years it has been accepted pro- vation of erythrocytes, it was decided to
cedure to collect specimens for coagulation use 3.8% citrate solution in the experideterminations into a solution of an anti- ments described below. Two features of
coagulant, usually sodium oxalate or sodium citrate make it preferable to sodium
sodium citrate. T h e most common ratio oxalate as an anticoagulant, despite the fact
used has been one part anticoagulant solu- that the latter is being used by about 20%
tion to nine parts whole blood. T h e ration- of survey participants: (1) factor V is better
ale for the dilution ratios and the concen- preserved in citrate; (2) calcium ion neutrations of anticoagulants used has not tralization is more rapid with citrate than
generally been well documented. Despite with oxalate. 1
this, several formulas have become tradiT h e remaining variable that needs to
tional and, in general, are no longer be standardized to assure optimal calcium
questioned.
ion concentration is the concentration of
calcium
in the recalcifying solutions in
A 3.2% (w./v.) solution of sodium citrate*
relation
to
the amount and concentration
is isosmolar, and might reasonably be
of
anticoagulant
solution used.
chosen as the optimal concentration. In
The studies reported in this paper were
contrast, the most widely used sodium citrate concentration is 3.8% (w./v.). This made, therefore, to examine several
concentration does cause shrinkage of variables that have been though to effect
erythrocytes of about 6%,4 which is quite the measurement of prothrombin time
and/or partial thromboplastin time. These
* Throughout this paper, the concentrations of variables include the hematocrit of the
sodium citrate solutions are given as the weight of
anhydrous sodium citrate (Na 3 C 6 H 5 0 7 ) per volume of specimen, the time elapsed between colwater. T h e most commonly used salt is the dihy- lection of the sample from the patient
drate form (Na 3 C 6 H 5 0 7 • 2H 2 0), for which appropriate corrections for the waters of hydration must and the test measurement, and the antibe made when preparing solutions.
coagulant concentration.
November 1975
593
VARIABLES IN COAGULATION TESTS
Effect of Hematocrit
Using standard plasmapheresis technics
(Fenwal system using N I H formula A anticoagulant), 600 ml. of citrated plasma and
300 ml. of erythrocytes were harvested
from two normal blood donors. Only the
plasma from the initial bleeding was retained, while the whole unit from the second bleeding was used. In order to remove
the anticoagulant as well as calcium, the
plasma was dialyzed in 8 1. Tris-saline
solution (0.154 M NaCl, .02 M Tris buffer,
pH 7.5) for three hours at room temperature, followed by dialysis overnight at 4 C.
in a plastic container, with agitation by a
magnetic stirrer. T h e plasma volume increased by approximately 4% during the
dialysis procedure. T h e calcium concentration of the dialyzed plasma was 1 mg.
per dl. and that of the dialysate, 0.75 mg.
per dl.
T h e packed erythrocytes were washed
five times in 5 volumes of Tris-saline
buffer for each volume of packed cells.
T h e cells were centrifuged at 4,000 x g for
10 minutes. They were then washed twice
in equal parts of dialyzed plasma and centrifuged for 20 minutes at 4,000 x g. T h e
packed erythrocytes were adjusted to a
hematocrit of approximately 98% by discarding the supernatant washings.
Four and a half milliliter specimens
with hematocrits ranging from 69 to 5%
were p r e p a r e d by adding increasing
amounts of dialyzed plasma to the washed
packed erythrocytes. One half milliliter of
3.8% (w./v.) sodium citrate anticoagulant
solution (which also contained an appropriate amount of 0.25 M CaCl2) was added
to give a final plasma calcium concentration of approximately 10 mg. per dl.
(Table 1). Thus, a series of simulated blood
specimens with a broad range of hematocrits was prepared, to which was added a
volume of anticoagulant solution usually
used in clinical laboratories. These specimens therefore contained increasing amounts
of citrate per volume of plasma as the
hematocrit levels increased. Calcium concentrations, in contrast, were similar over
the entire range of hematocrits.
A second set of simulated anticoagulated blood specimens was prepared in a
similar manner, with the exception of the
final citrate concentration. At a hematocrit
of 40%, with the addition of 0.5 ml. of
3.8% citrate solution, the final citrate concentration in the plasma used for testing
Table 2. Effects of Hematocrit on Prothrombin Time and Partial Thromboplastin Time
(Effects of Proposed Citrate Anticoagulant Concentrations)
Proposed Method
Usual Method
Measured
Hematocrit
++
++
(%)
Ca
(mg. per dl.)
PT
(Sec.)
PTT
(Sec.)
Ca
(mg. per dl.)
PT
(Sec.)
PTT
(Sec.)
4
10
16
20
24
29
35
40
44
49
54
59
64
69
8.0
11.0
9.8
11.6
8.4
9.4
10.6
8.6
9.8
8.6
8.0
8.0
8.8
9.0
13.0
13.0
13.0
13.0
13.0
13.0
13.2
13.5
13.5
13.5
14.2
14.5
15.7
19.0
29.5
31.9
32.6
31.5
30.9
31.7
33.6
36.1
34.6
39.7
39.2
45.9
67.7
79.9
10.4
8.0
9.0
9.2
9.0
9.2
9.6
8.8
8.4
13.0
7.8
8.0
8.4
7.8
12.0
12.7
13.0
13.2
13.2
13.2
13.5
13.0
13.5
13.0
13.0
12.7
13.2
12.7
31.3
32.9
31.7
31.7
32.3
29.9
32.8
31.4
32.4
33.4
35.8
31.9
34.4
38.1
594
KOEPKE, RODGERS, AND OLLIVIER
Ul
??
i
20
z tn
m u
18
o
16
S
(se
i-
cc
i
h-
o
IT
D.
12
14
80 r
W
5
i=
z
70-
3
60-
Q. _
O in
CD u
5 0)
in
O —
|
50-
3C.L I 9 ° I
5 10 20
1
30
HEMATOCRIT
1
40
1
50
1 1
6 0 70
(percent)
FIG. 1. Effects of hematocrit on prothrombin
time a n d partial thromboplastin time. Note the
marked increases in P T and P T T with routine
specimen collection procedures, i.e., 0.5 ml. 3.8%
sodium citrate added to 4.5 ml. of blood (solid
circles). This effect was obviated when the final sodium
citrate concentration of plasma was adjusted to
0.59% (open circles). Calcium concentrations of all
specimens adjusted to 10.6 ± .3 mg per dl. (mean
± S.D.).
A.J.C.P. —Vol. 64
studies a second normal d o n o r was
used as in the initial experiments described above, with the single exception
that the relative volume of 3.8% sodium
citrate anticoagulant was decreased. Here,
0.25 ml. of 3.8% citrate solution was added
to 4.75 ml. of whole blood specimens with
hematocrits ranging from 5 to 69%. The
citrate concentrations varied as in the
initial studies, increasing as the hematocrits
increased.
In all three studies, after thorough
mixing, the blood specimens were centrifuged at 740 Xg for 15 minutes. T h e
supernatant plasma was removed and
prothrombin times [Thromboplastin-C
(Dade)] and activated thromboplastin times
[Cephaloplastin (Dade)] were measured in
duplicate using a Fibrometer. Uniform
incubation times and calcium chloride concentrations were used in the recalcification
procedures, carefully following the manufacturer's instructions. In a separate set of
studies, a more concentrated calcium
chloride solution was used for recalcification. Final plasma calcium concentrations on all specimens were measured
using a Perkin-Elmer atomic absorption
spectrophotometer.
Both prothrombin times and activated
partial thromboplastin times were significantly prolonged in specimens with hematocrits above 60% collected with 0.5 ml.
of 3.8% sodium citrate anticoagulant in
4.5 ml. of whole blood. T h e coagulation
times progressively increased with inwas calculated to be 0.59%. This series of creasing hematocrit values (Fig. 1). When
specimens differed from the first series in the citrate concentration was adjusted to a
that these specimens were adjusted to a constant 0.59% citrate (based on 3.8%
uniform citrate concentration, i.e., 0.59%, solution at a hematocrit of 40%) in the
over the entire range of hematocrits. Cal- final plasma volume, the prothrombin
cium concentrations again were adjusted times remained constant at all hematocrit
to similar levels over the entire range of values. T h e partial thromboplastin time
increased only minimally over the entire
hematocrit values (Table 1).
An additional set of experiments was car- test range. By doubling the calcium conried out to assess whether decreasing the centration in the recalcification procevolume (and final concentrations) of citrate dures, coagulation times were longer at all
might be appropriate for routine use in hematocrit values, but did not change
clinical laboratories (Table 2). I n these significantly at higher hematocrit values.
November 1975
VARIABLES IN COAGULATION TESTS
595
With decreased volumes of 3.8% sodium
citrate (0.25 ml. vs. 0.5 ml. in a final volume of 5 ml.) the prothrombin times did
not lengthen at the higher hematocrit
values (see Fig. 2 and Table 2), as had occurred when larger volumes of citrate were
used (Fig. 1). Again, a minimal increase in
the partial thromboplastin time occurred.
There was no significant change in coagulation times at low (down to 5%) hematocrit
values.
Effect of Decreased Sodium Citrate
Anticoagulant Levels
Appropriate volumes of 3.8% (w./v.)
sodium citrate solution were removed from
Vacutainer tubes (3206 W) to give final
anticoagulant volumes of 0.25, 0.20, 0.175,
0.15, 0.125, 0.10, and 0.075 ml. Appropriate volumes of whole blood from a normal healthy subject were added to the
citrate solutions to make final volumes of
5 ml. T h e specimens were thoroughly
mixed immediately after collection. T h e
tubes containing 0.15 ml. or less of 3.8%
citrate solution clotted before centrifugation and were therefore not studied
further. Prothrombin time and partial
thromboplastin time tests were performed
as outlined above on the remaining specimens. After about 0.5 ml. of plasma had
been removed to perform these determinations, each tube was restoppered and allowed to incubate at room temperature
(about 25 C ) . After four hours a clot had
formed in the tube containing 0.175 ml.
of citrate solution.
Coagulation studies were repeated on
the remaining unclotted specimens. T h e
prothrombin times and partial thromboplastin times of all of the unclotted specimens were essentially the same in the
Vacutainer tubes containing 0.20, 0.25
and 0.50 ml. of 3.8% citrate solution.
5 10
20
30
HEMATOCRIT
40
50
60
70
(percent)
FIG. 2. Effects of hematocrit on prothrombin
time and partial thromboplastin time. In contrast to
data in Figure 1, the increase in times with increasing hematocrits using 0.5 ml. 3.8% sodium citrate
added to 4.5 ml. of blood (solid circles) was obviated
by decreasing the volume of 3.8% sodium citrate to
0.25 ml. added to 4.75 ml. of blood (open circles).
Calcium concentrations of all specimens adjusted to
9.3 ± 1.35 mg. per dl. (mean ± S.D.).
Vacutainer tubes (3206 W) from each often
healthy subjects (eight women and two
men). All specimens were centrifuged at
room temperature at 740 x g for 10 minutes. Prothrombin and partial thromboplastin times were determined as quickly as
possible after venipuncture, then again
after 2, 4, 6, and 24 hours of incubation
at room temperature. Separately collected, previously unopened specimens
Stability of Blood Specimens during
were measured at each interval. Each
Incubation at Room Temperature
specimen was tested using Dade reagents
Five 4.5-ml. blood specimens were with a Fibrometer, Ortho reagents with a
drawn into standard 3.8% sodium citrate Fibrometer, General Diagnostics reagents
596
KOEPKE, RODGERS, AND OLLIVIER
A.J.C.P. —Vol. 64
with a Coag-A-Mate instrument, and Hyland reagents with a Clotek instrument.
Each test was run according to the directions supplied by the manufacturer.
No significant change in either prothrombin time or partial thromboplastin
time occurred during storage of unopened
citrated specimens at room temperature
for as long as 6 hours. After 24 hours the
prothrombin time remained stable; however, a 1 0 - 1 5 % lengthening of the partial
thromboplastin time was seen after 24
hours of storage at room temperature.
Similar results were obtained with all four
coagulation systems in relation to their
respective normal values.
during an ongoing pilot study of a variety
of patients, including patients with polycythemia as well as anemia.
It is tempting to use increased amounts
of calcium for recalcification in specimens
with high hematocrit values in order to
overcome the apparent excessive amount
of citrate anticoagulant and thus avoid
changing presently popular collection
procedures. However, in checking this
possibility it was found that coagulation
times were increased significantly. Lovelock and Porterfield 5 had studied this problem many years ago a n d showed that
higher concentrations of calcium ions, in
fact, inhibited coagulation. Therefore, the
proposed decrease of citrate anticoagulant
seems to be the appropriate way to corDiscussion
rect the present erroneous collection proThese studies were undertaken to in- cedures for coagulation testing.
vestigate systematically some of the variT h e documentation of the stability of
ables that affect laboratory measurement
of prothrombin time and partial thrombo- citrated plasma for P T and P T T determiplastin time. Certain concentrations a n d nations if the specimen tubes are unopened
dilutions of anticoagulants have become should allow for improved laboratory
entrenched in laboratory practice with- blood collection procedures. T h e stability
out being critically evaluated. Likewise, a was evident in normal patient specimens
certain folklore has evolved which has as well as in a limited number of studies of
led to the widespread belief that specimens specimens from patients receiving Coufor routine coagulation studies require madin a n d samples with heparin antispecial p r e c a u t i o n s , including rapid coagulation. These studies d o not in any
chilling of specimens and special handling. way invalidate the known instability of
Results of our studies indicate that half separated plasma specimens when kept at
of the usual amount of 3.8% sodium room temperature for any length of time.
citrate anticoagulant solution is entirely
adequate for routine use, even when the
hematocrit is as low as 5%. T h e special advantage of this decreased amount of
citrate is the avoidance of the marked
spurious elevation of the P T and P T T so
often found in specimens with elevated
hematocrits, such as those from polycythemic patients. T h e use of lesser
amounts of anticoagulant has not been associated with any apparent disadvantage
References
1. Biggs R, MacFarlane RG: H u m a n Blood
Coagulation and its Disorder. Third edition.
Philadelphia, F. A. Davis, 1962, p. 136
2. Koepke JA: T h e 1969 survey of prothrombin
time. Am J Clin Pathol 54:502-507, 1970
3. Koepke JA: T h e partial thromboplastin time in
the CAP survey program. Am J Clin Pathol
63 (Suppl):990-994, 1975
4. Koepke JA: Personal observation
5. Lovelock JE, Porterfield, BM: Blood clotting:
T h e function of electrolytes and calcium.
Biochemistry 50:415-420, 1952