| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1
Texas Biotechnology Corporation, 7000 Fannin Street, Suite 1920, Houston, TX 77030, and
2
Department of Pathology, Stanford University Medical Center, Stanford, CA 94305;
a address correspondence to this author at: 3912 Marlowe St., Houston, TX 77005
Oral anticoagulants such as warfarin induce depletion of vitamin K-dependent coagulation factors, causing prolongation of the prothrombin time (PT) and the PT-International Normalized Ratio (INR). The available thromboplastins used for measuring the PT vary in their sensitivity to coagulation factor depletion, and each is assigned an International Sensitivity Index (ISI) that is used when calculating the INR. Traditional monitoring of oral anticoagulant therapy using the INR is confounded during concurrent therapy with direct thrombin inhibitors, such as argatroban, that also prolong the INR (1)(2)(3). Although alternative monitoring methods have been suggested for use in this setting (2)(3), guidelines allowing for refined interpretation of the INR would also be useful. To facilitate the development of such guidelines, we have characterized in vitro the differential effects of warfarin and argatroban, as well as the choice of thromboplastin used, on the INR during anticoagulation with both agents.
For this study, plasma specimens were obtained from 10 healthy donors (laboratory personnel) and 35 patients on warfarin therapy, in accordance with policies of the institution's responsible committee. Blood was collected into evacuated tubes (38 g/L sodium citrate), and plasma was prepared by the centrifugation of the blood within 30 min of collection. The only selection criterion for the patient group was that each was on warfarin therapy, and both inpatients and outpatients were included. In addition, four plasma pools (each prepared from at least 10 individuals) were obtained from Harris Laboratories. One plasma pool was from healthy donors, and three pools were from patients on warfarin therapy exhibiting generally similar levels of oral anticoagulation. Plasma specimens were stored at -30 °C for up to 2 weeks and thawed rapidly immediately before testing.
The PT of each individual or pooled plasma specimen was determined in
the absence and presence of argatroban
(NOVASTAN®; Texas Biotechnology Corp. and
SmithKline Beecham Pharmaceuticals) added in vitro at three
different, clinically-relevant concentrations (0.5, 1.0, and 2.0
µmol/L, final concentration). The argatroban concentrations were
selected to emulate those typically attained with infusion doses of 1,
2.5, and 5
µg · kg-1 · min-1,
respectively (4)(5). The PT assays on the individual
plasma specimens were performed using Thromborel S (ISI = 1.11;
Behring) on a Fibrintimer A (Behring). The PT assays on the pools were
performed using Thromborel S (different lot, ISI = 1.05),
Thromboplastin C-Plus (ISI = 1.92, Dade), and Simplastin Excel
(ISI = 2.01, Organon Teknika) on a Coag-A-Mate XC (Organon
Teknika). For each PT determined, the INR was calculated (6)
using the thromboplastin-specific ISI and the following equation:
 |
We performed analysis of variance on INR values from the plasma pools, using a general linear model. For all paired specimens, INR values in the presence of argatroban ("combined INR") were regressed on INR values in the absence of argatroban ("warfarin INR"). Data were logarithmically transformed to adjust for heteroscedasticity. Six regressions were evaluated, one for each combination of argatroban concentration (0.5, 1.0, and 2.0 µmol/L) and thromboplastin sensitivity (ISI of ~1 or ~2). For each combination, the data were fitted using a cubic polynomial model, and sequential tests of the regression coefficients indicated that only the linear effect was significant (i.e., the quadratic and cubic effects were nonsignificant). Tests for common intercept and common slope were performed to evaluate the comparability of the two reagent lots used, with an ISI of ~1, and of the two different reagents used, with ISI values of ~2. We used the log-transformed data to estimate 95% prediction intervals for the combined INR at each warfarin INR for each of the six combinations of argatroban concentration and thromboplastin sensitivity. The error in the warfarin INR predicted from a combined INR was estimated as the half-width of the inverse prediction interval.
Fig. 1
shows the arithmetic linear relationships between the paired
INR values measured in the absence or presence of argatroban, using the
thromboplastin with an ISI of ~1 (Fig. 1A
) and the two
thromboplastins with ISI values of ~2 (Fig. 1B
). Tests of common
intercept and common slope indicated that the two different lots of the
low ISI thromboplastin (ISI values of 1.05 and 1.11) were comparable,
as were the two different thromboplastins with ISI values of 1.92 and
2.01. Irrespective of the choice of thromboplastin, the combined INR
increased linearly with the warfarin INR at each plasma argatroban
concentration tested (r 
0.95). The sensitivity of the
combined INR to the warfarin INR (i.e., the slope of the regression
line) increased significantly (P <0.0001) with plasma
argatroban concentration over the range evaluated. In addition,
the sensitivity of the combined INR to argatroban increased when the
ISI of the thromboplastin increased approximately twofold (ISI values
of ~1 vs ~2; P <0.0001).
|
The variance about the regression lines increased with INR, but was homogeneous when the data were logarithmically transformed. On the basis of the inverse 95% prediction intervals of the log-transformed data, the error in the warfarin INR predicted from a combined INR tended to increase with plasma argatroban concentration for the thromboplastin with ISI of ~1, but not the for the thromboplastin with ISI ~2. The maximum errors (detransformed to the arithmetic scale) in the predicted warfarin INR for a combined INR of 5 were ± 0.4, ±0.5, and ± 0.6 INR units, respectively, for 0.5, 1.0, and 2.0 µmol/L argatroban, respectively. At the same respective argatroban concentrations, the maximum prediction errors for a combined INR of 7 were ± 0.5, ±0.7, and ± 0.9 INR units, respectively.
We concluded that the combined effects exerted by warfarin and
argatroban anticoagulation on the INR are sensitive to the degree of
warfarin-induced factor depletion, the plasma argatroban concentration,
and the thromboplastin used. Specifically, the use of a low-ISI
thromboplastin minimized the combined effects. In addition, these
combined effects can be modeled linearly (Fig. 1
). It may be possible
to exploit this model to predict the contribution of warfarin to the
INR during concurrent argatroban therapy, given knowledge of the
combined INR, the plasma argatroban concentration, and the
thromboplastin used. Although the relationship between the differential
contribution of warfarin to the combined INR and bleeding risk during
concurrent argatroban therapy remains to be established, a predicted
warfarin INR could be useful for approximating the degree of
anticoagulation expected upon cessation of argatroban or perhaps for
adjusting warfarin dosage during argatroban therapy. However, on the
basis of the estimated error in the predicted warfarin INR, this model
may have limited clinical utility for specimens containing >1.0
µmol/L argatroban (or hence for argatroban infusion doses greater
than ~2.5
µg · kg-1 · min-1).
For combined INR values of 5 and 7, for example, the maximum error in
the predicted INR was generally acceptable (i.e., less than or equal
to ± 0.5 units) at 0.5 µmol/L argatroban, yet it became less
acceptable (i.e., up to ± 0.9 units) at 2.0 µmol/L argatroban.
The development of a similar model based on argatroban dosage
rather than concentration may be feasible because these parameters are
linearly related (4) and would perhaps further facilitate
practical application of the model.
The differential effects of warfarin and argatroban on the INR (Fig. 1
)
were evaluated using the INR values of healthy donors and patients
(both inpatients and outpatients), as well as pools of these
individuals, determined in the presence and absence of argatroban. The
variability in the model therefore would be expected to reflect, at
least in part, interindividual variability in factors other than
warfarin and argatroban that could possibly influence the INR, such
fibrinogen and fibrin degradation products. Another factor, if present,
that could potentially influence the INR would be an argatroban
metabolite. However, argatroban has only one metabolite that can be
detected in plasma, and it exerts no significant anticoagulant effects
at clinically relevant concentrations (7).
In clinical laboratories, many different combinations of thromboplastin and clot detection instruments are used for determining the INR (8). In this study, four different combinations of thromboplastins and instruments were used for determining INR values, and for each combination, the combined effects exerted by argatroban and warfarin on the INR could be modeled linearly. The thromboplastins used were of plain composition (i.e., not combined with adsorbed plasma nor of recombinant source), with ISI values of ~1 and ~ 2. One could speculate that for a given argatroban concentration, the sensitivity of the combined INR to the warfarin INR for a thromboplastin with an intermediate ISI value (e.g., 1.3 or 1.8) would fall between those for thromboplastins with ISI values of 1 and 2. Future confirmatory studies that evaluate other combinations of instruments and thromboplastins, including perhaps those of different ISI values and of different type (combined or recombinant), may be valuable.
The extent to which these conclusions apply to other oral anticoagulants and antithrombin agents is not yet clear. However, on the basis of these promising in vitro results, clinical studies have been initiated to define this model more fully and to investigate further the utility of the INR, with refined interpretation, for monitoring warfarin during concurrent argatroban therapy.
Footnotes
fax 713-432-7498, e-mail mhursting{at}sprintmail.com
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  J. M. Walenga, A. F. Drenth, M. Mayuga, D. A. Hoppensteadt, M. Prechel, S. Harder, H. Watanabe, M. Osakabe, and H.-K. Breddin Transition From Argatroban to Oral Anticoagulation With Phenprocoumon or Acenocoumarol: Effect on Coagulation Factor Testing Clinical and Applied Thrombosis/Hemostasis, July 1, 2008; 14(3): 325 - 331. [Abstract] [PDF]
|
|
|
|
 |
|
 D. W Kubiak, P. M Szumita, and J. R Fanikos Ann. Pharmacother., November 1, 2005; 39(11): 1956 - 1957. [Full Text] [PDF]
|
|
|
|
|
|
S. A Spinler and W. Dager Comment: extensive prolongation of aPTT with argatroban in an elderly patient with improving renal function, normal hepatic enzymes, and metastatic lung cancer Ann. Pharmacother., November 1, 2005; 39(11): 1955 - 1956. [Full Text] [PDF]
|
|
|
|
|
|
M. J. Hursting, B. E. Lewis, and D. E. Macfarlane Transitioning from Argatroban to Warfarin Therapy in Patients with Heparin-induced Thrombocytopenia Clinical and Applied Thrombosis/Hemostasis, July 1, 2005; 11(3): 279 - 287. [Abstract] [PDF]
|
|
|
|
 |
|
 I.-K. Jang and M. J. Hursting When Heparins Promote Thrombosis: Review of Heparin-Induced Thrombocytopenia Circulation, May 24, 2005; 111(20): 2671 - 2683. [Full Text] [PDF]
|
|
|
|
|
|
J. L Stephens, J. M Koerber, J. C Mattson, and M. A Smythe Effect of Lepirudin on the International Normalized Ratio Ann. Pharmacother., January 1, 2005; 39(1): 28 - 31. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Smythe, W. E. Dager, and N. M. Patel Managing Complications of Anticoagulant Therapy Journal of Pharmacy Practice, October 1, 2004; 17(5): 327 - 346. [Abstract] [PDF]
|
|
|
|
|
|
W. E Dager and R. H White Low-Molecular-Weight Heparin-Induced Thrombocytopenia in a Child Ann. Pharmacother., February 1, 2004; 38(2): 247 - 250. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. S Trask, R. C Gosselin, J. A Diaz, and W. E Dager Warfarin Initiation and Monitoring with Clotting Factors II, VII, and X Ann. Pharmacother., February 1, 2004; 38(2): 251 - 256. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. E Dager and R. H White Argatroban for Heparin-Induced Thrombocytopenia in Hepato-Renal Failure and CVVHD Ann. Pharmacother., September 1, 2003; 37(9): 1232 - 1236. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. M. Alving How I treat heparin-induced thrombocytopenia and thrombosis Blood, January 1, 2003; 101(1): 31 - 37. [Full Text] [PDF]
|
|
|
|
 |
|
 T. Fenyvesi, I. Joerg, and J. Harenberg Influence of Lepirudin, Argatroban, and Melagatran on Prothrombin Time and Additional Effect of Oral Anticoagulation Clin. Chem., October 1, 2002; 48(10): 1791 - 1794. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
