Coagulation Changes in Individuals With Sickle Cell Trait

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

American Journal of Hematology 69:89±94 (2002)

DOI 10.1002/ajh.10021

Coagulation Changes in Individuals With Sickle Cell Trait


M.P. Westerman,1,3* D. Green,2 A. Gilman-Sachs,3 K. Beaman,3 S. Freels,4 L. Boggio,2 S. Allen,1
R. Schlegel,5 and P. Williamson6
1
Department of Medicine, Mount Sinai Hospital Medical Center, Chicago, Illinois
2
Department of Medicine, Northwestern University School of Medicine, Chicago, Illinois
3
Department of Immunology, University of Health Sciences/Chicago Medical School, North Chicago, Illinois
4
Department of Biostatistics, University of Illinois School of Public Health, Chicago, Illinois
5
Department of Molecular and Cell Biology, Pennsylvania State University, University Park, Pennsylvania
6
Department of Biology, Amherst College, Amherst, Massachusetts

Sickle cell disorders, such as Hb SS and Hb SC, are associated with a hypercoagulable
state that may contribute to the vaso-occlusive episodes observed in the disorders. To
what extent increased coagulation activity occurs in individuals with sickle cell trait has
had limited study. Because such information may help clarify clinical and pathologic
®ndings that may occur in these individuals and may be useful in clarifying the hyper-
coagulable state in sickle cell disease, we have examined individuals with Hb AS to
determine the extent that increased coagulation activity does occur. We measured
d-dimers, thrombin±antithrombin (TAT) complexes, prothrombin fragment 1.2 (F1.2),
absolute blood monocyte levels, proteins C and S, and isotypes of antiphospholipid
antibodies in individuals with Hb AS and in matched controls (Hb AA). Results showed
that d-dimers, TAT, and F1.2 were increased signi®cantly above normal levels. Absolute
blood monocyte levels were increased. The d-dimers, TAT, F1.2, and monocyte counts
showed signi®cant increasing trends through groups of increasing severity (Hb AA, Hb
AS, Hb SC, and Hb SS). Our study shows that individuals with Hb AS have increased
coagulation activity, with d-dimers, TAT, and F1.2 being consistent indicators. The
measures of coagulation activity in Hb AS are lower than in patients with Hb SC and Hb
SS disease. These results extend our previous observation that the degree of coagulation
activation parallels the degree of disease severity among sickle cell genotypes. The
®ndings suggest that monocytosis, with the possible expression of monocyte-derived
tissue factor, and the associated hypercoagulable state are driven by disease severity.
Am. J. Hematol. 69:89±94, 2002. ã 2002 Wiley-Liss, Inc.

Key words: sickle cell trait; coagulation; coagulation activity; hypercoagulability

INTRODUCTION patients with Hb SS and Hb SC disease, we have


made coagulation measurements in individuals with
Although sickle cell trait (Hb AS) is a benign dis- the trait and have compared results to controls,
order, there is an increased risk for several abnor- matched for age, gender, and race, and to previously
malities, including splenic infarction at high altitudes, described groups of patients with Hb SS and Hb SC
hyposthenuria, hematuria, bacteriuria in women, disease [3].
bacteriuria and pyelonephritis in pregnancy, sudden
death following exertion, pulmonary embolism, and
glaucoma due to anterior chamber bleeding [1]. Pre- METHODS
vious studies have shown that a hypercoagulable state Twenty-three African-Americans with sickle cell
is present in patients with Hb SS and Hb SC [2,3] and trait (Hb AS) and 23 controls (Hb AA), who were
in sickle cell trait [2] and that the changes in coagu-
lation activity in Hb SS and Hb SC disease re¯ect the
degree of disease severity of the hemoglobin disorder *Correspondence to: M.P. Westerman, M.D., Hematology/On-
[3]. To further examine the extent to which increased cology Unit, Mount Sinai Hospital Medical Center, 15th at Cali-
fornia Avenue, Chicago, IL 60608. E-mail: [email protected]
coagulation activity occurs in individuals with sickle
cell trait and to what degree the changes are compa- Received for publication 20 January 2001; Accepted 15 August
rable to the coagulation abnormalities observed in 2001
ã 2002 Wiley-Liss, Inc.
90 Westerman et al.

matched for age, gender, and race, were examined. Protein S activity was measured by a clot-based
The individuals were 20 73 years of age and included analysis of dilute plasma with bu€er (1:5) and prein-
19 women and 4 men. The controls and patients were cubated with added activators and protein S de®cient
randomly selected and consisted of laboratory and plasma (Bioclot Protein S kits, Biopool, Ventura, CA)
non-laboratory personnel. Technicians performing [6]. Protein S antigen (total, free) was analyzed by
the laboratory assays were blinded as to the origin of EID on antibody-coated agarose plates (Assera-plate
all samples. The study was approved by the Institu- PS antigen kindly supplied by Diagnostica Stago).
tional Review Board of Mount Sinai Hospital, and Total protein S antigen was determined by applying
informed consent was obtained from the participants. native plasma to the plate. Free protein S (unbound
to C4b-BP of complement) was also determined by
EID after separating the free from the bound protein
S. This was accomplished by cold precipitation of the
Assays of Coagulation Activation bound protein S after mixing plasma (13:1) with 25%
Measures of coagulation activation included d-di- PEG (polyethylene glycol), centrifuging and separat-
mers, thrombin antithrombin complexes (TAT), and ing the supernatant as free protein S.
prothrombin fragment 1.2 (F1.2). The d-dimers were Absolute blood monocyte counts were calculated
assayed by ELISA (American Bioproducts Co.; Di- from the percentage of monocytes present in total
agnostica Stago, Asnieres-sur-Seine, France), TAT leukocyte counts. Urine osmolality was determined
was determined by ELISA (Dade Behring, Chicago, by freezing point depression (Advanced Instruments,
IL), and F1.2 was assayed by ELISA (Dade Behring), Norwood, MA) on overnight fasting, ®rst morning
as previously described [3]. specimens.

Antiphospholipid Antibody Assay Statistical Methods


Antiphospholipid antibodies (phosphatidylserine The distributions of clotting variables (d-dimers,
[PS], cardiolipin [CL], phosphatidylethanolamine TAT, F1.2) and monocyte counts within disease
[PE], phosphatidylglycerol [PG], and phosphatidyli- groups tended to be skewed with large outliers to the
nositol [PI]) were measured by an ELISA as previ- right, and most showed evidence of extreme depar-
ously described [4]. Cut-o€ values were determined by tures from normality according to the Shapiro Wilk
using sera from 46 non-pregnant controls between the test. Due to the evidence of non-normality, together
ages of 20 and 50 years. An absorbance equal to or with the relatively small sample sizes on each group,
greater than 2 standard deviations above the mean for we used non-parametric methods to compare medians
that phospholipid and antibody isotype was consid- rather than means. Correlations are based on rankings
ered positive. In some instances multiples of the mean rather than original values.
have been used with essentially the same results. b2- The distribution of the measures of coagulation
GPI was a necessary co-factor and was present in the activity were described graphically and summarized
fetal calf serum used as the blocker. with median values and inter-quartile ranges (IQR).
Median values of each measure (d-dimers, TAT,
F1.2) were each compared between the individuals
with Hb AS and their matched controls using sign
Proteins C and S Assays tests. Median values were also compared between the
Proteins C (activity, antigen) and S (activity, total, individuals with Hb AS and patients with Hb SC
free antigen) were measured. The plasma was tested disease, between the individuals with Hb AS and
immediately for activity measurements and the re- patients with Hb SS disease and between patients with
mainder was quick frozen at 80°C until the time of Hb SC and patients with Hb SS [3] using rank sum
the antigenic determinations. Appropriate standards tests. Spearman correlation coecients were used to
and controls were performed with all assays. examine the association between each of the coagu-
Protein C activity was determined by chromogenic lation measures and urine osmolality.
analysis using S2366 (Diapharm Chromogenic, Correlations and partial correlations were used to
MoÈlndal, Sweden) as the chromophore and Protac analyze associations between d-dimers, monocyte
(American Diagnostica, Greenwich, CT) as the pro- counts, and disease severity. Disease severity was an-
tein C activator. Protein C antigen was determined by alyzed as a score representing the degree of increasing
electroimmunodi€usion (EID) (Laurell rocket elec- clinical severity in the various groups: 0 ˆ Hb AA,
trophoresis). Agarose antibody plates were obtained 1 ˆ Hb AS, 2:5 ˆ Hb SC, 3 ˆ Hb SS. Thus the vari-
from Helena Laboratories (Beaumont, TX) [5]. able, disease severity, would re¯ect the e€ect of an
Coagulation Changes in Sickle Cell Trait 91

average trend or the e€ect of a unit increase in the TABLE I. Clinical Data on Individuals With Sickle Cell Trait
assigned score. Correlations were computed between Hb HbS MCV Urine osmolality
d-dimers and monocyte counts, both unadjusted and Age Gender (g/dl) (%) (¯) (mOsm/kg H2O)
adjusted for group scores; between group scores and
monocyte counts, both unadjusted and adjusted for d- 1. 30 F 13.5 42 88 546
2. 34 F 13.9 42 89.5 592
dimers; and, between group scores and d-dimers, both 3. 31 F 13.3 37 79.4 427
unadjusted and adjusted for monocyte counts [7]. 4. 46 F 13.9 42 93.4 247
5. 46 F 13.3 41 93.9 238
6. 28 M 15.1 42 94.4 801
RESULTS 7. 20 M 14.7 43 91 778
8. 72 M 14.5 40 87.4 469
Clinical data are shown in Table I. All participants 9. 27 F 12.8 26 72 787
10. 43 F 12.4 41 87.5 462
in the study were healthy with no evidence of disease. 11. 56 F 13.2 40 85 467
Two individuals (#13, #22) had a past history of gross 12. 50 M 12.9 41 87.5 439
hematuria on one occasion. None of the individuals 13. 58 F 13.5 40 89.5 457
with sickle cell trait had present or past history of 14. 43 F 10.8 41 92 412
pulmonary embolism or leg ulcers. Ten individuals 15. 37 F 11.1 42 82.7 388
16. 28 F 11.4 43 93 NDa
(#14 #23) had moderately decreased hemoglobin 17. 50 F 11.7 40 94.6 817
levels (10.4 11.8 g/dl). The individuals with de- 18. 47 F 10.8 38 88.6 338
creased hemoglobin levels had no history of preg- 19. 72 F 11.8 40 79 466
nancy for at least 2 years, no clinical or laboratory 20. 25 F 11.7 39 80.9 472
evidence of iron de®ciency, and no reported use of 21. 73 F 11.4 42 83 432
22. 37 F 10.4 31 81 472
oral contraceptives during the period of the study. 23. 30 F 10.5 39 81 674
The decrease in hemoglobin levels in this group may
a
be related to several factors such as co-existent ND, not done.
a-thalassemia [8], aging, race, or gender [9].
Median d-dimer, TAT, and F1.2 levels were sig- Absolute blood monocyte counts were lowest in the
ni®cantly increased in the individuals with sickle cell controls and were progressively higher in individuals
trait (Table II). There was no correlation between d- with Hb AS and in patients with Hb SC and Hb SS
dimer, TAT, and F1.2 and the hemoglobin concen- (controls, 398 [274] ´ 109 cells/l; median [IQR]; Hb
tration, whether normal or decreased. Similarly, the AS, 476 [301]; Hb SC, 613 [ 691]; Hb SS, 949 [867]).
levels of Hb S and urine osmolality did not correlate Pairwise comparisons between groups were not sig-
with the levels of coagulation activity. The d-dimer, ni®cant at P < 0:05. The plots in Fig. 2 show the
TAT, and F1.2 levels in individuals with Hb AS were group median values for monocyte counts plotted
less than those previously described in patients with against the group median values for d-dimers and
Hb SC and Hb SS disease (Table II, Fig. 1) [3]. re¯ect a near-linear relationship across the group

TABLE II. Measures of Coagulation Activation in Individuals With Sickle Cell Trait, Patients With Hb SC Disease, Patients With Hb
SS Disease, and in Controlsa

Sickle trait Control (sickle trait) Hb SC Hb SS Control (Hb SC, HbSS)

D-dimer (lg/ml) 0.46 (0.27) [n = 23] 0.30 (0.25) [n = 23] 0.81 (0.70) [n = 16] 1.26 (1.63) [n = 18] 0.2 (0.2) [n = 17]
P 0.0005*, 0.0013**; 0.0792b
0.0002***
TAT (ng/ml) 2.8 (2.5) [n = 23] 1.5 (1.0) [n = 23] 4.1 (3.7) [n = 15] 6.8 (8.1) [n = 18] 2.4 (2.35) [n = 18]
P 0.0072*, 0.0379**; 0.0203b
0.0003***
F1.2 (nm) 0.6 (0.5) [n = 23] 0.5 (0.5) [n = 23] 2.3 (1.8) [n = 15] 2.6 (1.3) [n = 18] 0.76 (0.76) [n = 17]
P 0.0266*, 0.0004**; NSb
0.0002***
a
Values are expressed as median (inter-quartile range); values for Hb SC and Hb SS were obtained from Ref. 3; NS, not signi®cant at <0.05 level;
n = number of subjects.
b
Hb SC v Hb SS.
*Sickle trait v control.
**Sickle trait v Hb SC.
***Sickle trait v Hb SS.
92 Westerman et al.

Fig. 1. Levels of d-dimers, TAT, and F1.2 in patients with Hb SS disease, Hb SC disease, in individuals with sickle cell trait
and in controls. (IÐ-I) Median values for Hb SS, Hb SC, and Hb AS. (OÐ-O) Median values for controls.

means, with the groups ordered in terms of increasing adjusting for monocyte counts, was r ˆ 0:78 (P <
disease severity. Plots of TAT and F1.2 did not show 0:0001, n ˆ 57). The correlation coecient for the
a similar trend. trend in monocyte counts by disease severity score,
Increasing trends in d-dimers and monocyte counts adjusting for d-dimers, was r ˆ 0:33 (P ˆ 0:013, n ˆ
computed from all disease groups combined showed 57). The association between d-dimers and monocyte
an apparent association between d-dimers and counts was not signi®cant after adjusting for group.
monocyte counts without adjusting for group (r ˆ Antiphospholipid antibodies, protein C (activity,
0:27, P ˆ 0:04, n ˆ 57). The correlation coecient for antigen), and protein S (activity, total, free antigen) in
the trend in d-dimers by disease severity score individuals with Hb AS were in the normal range
(0 ˆ Hb AA, 1 ˆ Hb AS, 2:5 ˆ Hb SC, 3 ˆ Hb SS), (results not shown).
Coagulation Changes in Sickle Cell Trait 93

types as disease severity increases from controls


through Hb AS, Hb SC disease, and Hb SS disease.
The partial correlation between disease severity
groups and monocyte counts, adjusting for d-dimers,
is signi®cant, whereas the partial correlation between
d-dimers and monocyte counts, adjusting for disease
severity groups, is not signi®cant. The ®nding sug-
gests that membership in a genotype has a stronger
independent relationship with monocyte counts than
do d-dimer levels and implies that the groups are
driving the relationship between absolute monocyte
levels and likely tissue factor expression and coagu-
lation activity.
The measures of coagulation activity appear to
re¯ect the levels of disease severity observed in the
groups, i.e., in Hb SS disease, Hb SC disease, and Hb
AS. Levels of d-dimers, TAT, and F1.2 decline in
parallel with the decreasing degrees of disease severity
observed in the genotypes with TAT and d-dimer
levels being the most consistent measures of increased
Fig. 2. Comparison of median absolute monocyte counts
to median d-dimer levels in patients with Hb SS disease,
coagulation activity. The extent to which the di€er-
Hb SC disease, in individuals with sickle cell trait, and in ences in coagulation activity contribute to the di€er-
controls. ences in clinical and pathologic ®ndings in Hb SS
disease, Hb SC disease, and sickle cell trait is not
DISCUSSION clear. The ®ndings of a hypercoagulable state in the
individuals with Hb AS extend the results obtained
The study shows that the coagulation pathway in in our previous study [3] in which the degree of
individuals with sickle cell trait is activated, as indi- hypercoagulability re¯ected the di€erence in disease
cated by the signi®cantly increased levels of d-dimers, severity between Hb SS and Hb SC disease. The
TAT, and F1.2. The cause of the increased coagula- progressive decline in absolute monocyte counts with
tion activity includes several possibilities. Loss of the decreasing levels of group disease severity would
phospholipid asymmetry in the membranes of sickle be compatible with the ®ndings of a similar decline in
red blood cells [2,10 12] and in the membranes of the levels of coagulation activity associated with decrea-
vesicles shed from the cells [13 15], with exposure of sing group disease severity.
phosphatidylserine (PS) on the membrane surface, The ®ndings in our study di€er from those ob-
would contribute to hypercoagulability. Although the served in a previous study [2] in that d-dimers, which
loss of asymmetry, which is associated with deoxy- were signi®cantly altered in the present study, were
genation of sickle red cells and production of mem- not measured in the earlier study and the relation-
brane-shed vesicles, would be less likely to occur in ships between group severity and coagulation activity,
Hb AS red cells than in Hb SS or Hb SC red cells, loss although parallel in the previous study, did not
of asymmetry can occur because deoxygenation of red generally reach statistical signi®cance.
cells is observed in individuals with Hb AS [16]. In- Evaluation of the hypercoagulable state and its re-
creased absolute blood monocyte levels, which are lationship to the clinical and pathologic ®ndings in Hb
associated with increased expression of monocyte- AS have not been de®ned. Measures of coagulation
derived tissue factor in both Hb SS and Hb SC dis- activity during complications, such as hematuria,
ease [17], are also increased in individuals with Hb AS however, may be important. Although the total num-
and may contribute to the increase in coagulation ber of patients who have been examined in the present
activity [18]. Consequently, both the loss of phosp- study are relatively small, we consider that the ®ndings
holipid asymmetry in sickle red cells and the sickle red are relevant and may contribute to concepts of
cell-shed vesicles and the presence of monocytosis, pathogenesis and management. The ®ndings may also
with possible increased expression of monocyte-de- contribute to an understanding of the increased risk of
rived tissue factor, may contribute to the increase in unexplained sudden death in African-American re-
coagulation activity in individuals with Hb AS. cruits in the Armed Forces with Hb AS or to the more
It is of interest that the levels of d-dimers and blood recently described increased risk of pre-eclampsia ob-
monocytes show a near-linear trend across the geno- served in African-American women with Hb AS [19].
94 Westerman et al.
REFERENCES 10. Lubin B, Chiu D, Bastacky J, Roelofson B, van Deenan LLM.
Abnormalities in membrane phospholipid organization in sickled
1. Sears DA. Sickle cell trait. In: Embury SH, Hebbel RP, Mohandas erythrocytes. J Clin Invest 1981;67:1643 1649.
N, Steinberg MH, editors. Sickle cell disease. Basic principles and 11. Chiu D, Lubin B, Roelofsen B, van Deenen LLM. Sickled ery-
clinical practice. New York: Raven Press, Ltd; 1994. p 381 94. throcytes accelerate clotting activity in vitro: an e€ect of abnormal
2. Helley D, Girot R, Guillin M, Bezaud A. Sickle cell disease: re- membrane lipid asymmetry. Blood 1981;58:398 401.
lation between procoagulant activity of red blood cells from dif- 12. Tait JF, Gibson D. Measurement of membrane phospholipid
ferent phenotypes and in vivo blood coagulation activation. Br J asymmetry in normal and sickle cell erythrocytes by means of
Haematol 1997;99:268 272. annexin V binding. J Lab Clin Med 1994;123:741 748.
3. Westerman MP, Green D, Gilman-Sachs A, Beaman K, Freels S, 13. Westerman MP, Cole ER, Wu K. The e€ect of spicules obtained
Boggio L, Allen S, Zuckermen L, Schlegel R, Williamson P. An- from sickle red cells on clotting activity. Br J Haematol
tiphospholipid antibodies, proteins C and S, and coagulation 1984;56:557 562.
changes in sickle cell disease. J Lab Clin Med 1999;134:352 362. 14. Franck PH, Bevers EM, Lubin BH, et al. Uncoupling of the
4. Gilman-Sachs A, Lubinski J, Beer AE, Brend S, Beaman K. Pat- membrane skeleton from the lipid bilayer. The cause of accelerated
terns of antibody speci®cities. J Clin Lab Immunol 1991;35:83 88. phospholipid ¯ip-¯op leading to an enhanced procoagulant
5. Bertina RM, Broekmans AW, Krommenhock-van Es C, van activity of sickled cells. J Clin Invest. 1985;75:183 190.
Wijngaarden A. The use of a functional and immunologic assay 15. Choe HR, Schlegel R, Rubin E, Williamson P, Westerman MP.
for plasma protein C in the study of heterogenicity of congenital Alteration of red cell organization in sickle cell anaemia. Br J
protein C de®ciency. Thromb Haemost 1984;51:1 5. Haematol 1986;63:761 773.
6. Boyer-Neuman C, Bertina RM, Tripodi A, et al. Comparison of 16. Martin TW, Weisman IM, Zeballos RJ, Stephenson SR. Exercise
functional assays for protein S. Thromb Haemost 1993;70: and hypoxia increase sickling in venous blood from an exercising
946 950. limb in individuals with sickle cell trait. Am J Med 89;87:48 56.
7. Rosner B. Fundamentals of biostatistics. 4th edition. Belmont, CA: 17. Key NS, Slungaard A, Dandelet L, et al. Whole blood tissue factor
Duxberry Press; 1995. procoagulant activity is elevated in patients with sickle cell disease.
8. Steinberg MH, Embury SH. a-Thalassemia in blacks: genetic and Blood 1998;91:4216 4223.
clinical aspects and interactions with sickle hemoglobin genes. 18. Belcher JD, Marker PH, Weber JP, Hebbel RP, Vercellotti GM.
Blood 1986;68:985 990. Activated monocytes in sickle cell disease: potential role in the
9. Salive ME, Cornoni-Huntley J, Guralnik JM, Philips CL, Wallace activation of vascular endothelium and vasocclusion. Blood
RB, Ostfeld AM, Cohen HJ. Anemia and hemoglobin levels in 2000;96:2451 2459.
older persons: relationship with age, gender, and health status. 19. Larrabee KD, Monga M. Women with sickle trait are at increased
J Am Geriatr Soc 1992;40:489 496. risk for pre-eclampsia. Am J Obstet Gynecol 1997;2:425 428.

You might also like