Protein C and protein S levels can be accurately determined within 24 hours of diagnosis of acute venous thromboembolism

Clin. Lab. Haem. 2006, 28, 9 13 M. J. KOVACS*, J. KOVACS*, J. ANDERSON, M. A. RODGER, K. MACKINNON, P. S. WELLS Summary Keywords Protein C and protein S levels can be accurately determined within 24 hours of diagnosis of acute venous thromboembolism *Department of Medicine, London Health Sciences Centre, London, ON, Canada Department of Medicine, The Ottawa Hospital, Ottawa, ON, Canada àdepartment of Laboratory Medicine, London Health Sciences Centre, London, ON, Canada In the 50% of cases of acute idiopathic venous thromboembolism, laboratory testing for inherited causes is often performed. Most physicians are under the impression that assays for protein C and protein S should not be measured at the time of diagnosis because of a high false positive rate. We performed a prospective cohort study from two outpatient thromboembolism clinics on consecutive patients with an objectively confirmed diagnosis of first acute idiopathic venous thromboembolism. Assays for protein C and protein S were performed prior to the initiation of oral anticoagulation therapy and within 24 h of diagnosis of venous thromboembolism. Abnormal results were repeated when patients discontinued oral anticoagulant therapy. Of 253 patients tested for both protein C and protein S, 229 (91%; 95% confidence interval 87 94%) were negative and 484 of 508 (95%) tests were normal. Of the 24 initial abnormal results, 21 were repeated and 10 (48%; 95% confidence interval 26 70%) were still abnormal. Overall, 97.8% of initial protein C and protein S results were accurate. If protein C and protein S are measured at the time of diagnosis of acute venous thromboembolism, the majority of the results will be normal and false positives are uncommon. Diagnosis, venous thrombosis, thrombophilia, protein C, protein S Received 21 June 2005; accepted for publication 31 October 2005 Correspondence: Dr Michael J. Kovacs, London Health Sciences Centre, Victoria Hospital, 800 Commissioners Road East, London, ON N6A 4G5, Canada. Tel.: +519 685 8475; Fax: +519 685 8477; E-mail: michael.kovacs@lhsc.on.ca Ó 2006 Blackwell Publishing Ltd Venous thromboembolism is a common medical problem (Salzman & Hirsh, 1994; Silverstein et al., 1998). The etiology of an episode of venous thromboembolism includes transient risk factors (surgery, trauma, etc.), estrogen therapy and active malignancy (Weinmann & Salzman, 1994). However, approximately 50% of patients with venous thromboembolism have no apparent cause (Lane et al., 1996a, b; Bauer, 2001) These are termed idiopathic or unprovoked. For patients who have idiopathic venous thromboembolism, laboratory testing for inherited or acquired causes is often performed to help identify the etiology and possibly to help determine the optimal duration of oral anticoagulant therapy (Kearon, Crowther & Hirsh, 2000; Seligsohn & Lubetsky, 2001; Crowther & Kelton, 2003). A typical screen includes an assessment for inherited risk factors by means of quantitative assays (protein C, protein S, antithrombin) or genetic tests (Factor V Leiden or Prothrombin Gene 20210A; Bauer, 2001). Genetic testing is not influenced by oral anticoagulant therapy, but the quantitative assays for protein C and protein S are by virtue of their requirement for vitamin K catalyzed gamma carboxylation. It has also been reported that heparin interferes with the assessment of antithrombin levels. Furthermore, it has been suggested that the quantitative assays should not be measured at the time of diagnosis of acute venous thromboembolism. D Angelo et al. (1988) were the first to suggest that protein S levels decrease during acute deep vein thrombosis secondary to higher levels of C4bBP. Since then many others have repeated this warning (Lane et al., 1996a, b; Cumming & Shiach, 1999; Kearon et al., 2000; Bauer, 2001; Seligsohn & 9

10 Protein C and protein S can be accurately determined Lubetsky, 2001; Thomas, 2001). The perception is that protein C and/or protein S levels are falsely lowered presumably because of consumption and measurements should be performed at a later date. This requires the discontinuation of warfarin and possibly bridging therapy with anticoagulants such as low molecular weight heparin (Bauer, 2001). For those who need to remain on warfarin, this can be inconvenient and expensive. However, D Angelo s study reporting on only eight patients simply described lower mean protein S levels and did not indicate what proportion of results were below normal. Ideally, it would be simpler to measure protein C and protein S levels at the time of diagnosis of venous thromboembolism before warfarin is initiated. This study assesses the practice of measuring protein C and protein S levels at the time of diagnosis of venous thromboembolism. Materials and methods Patients We studied consecutive patients presenting to the outpatient thromboembolism clinics at two university hospitals with a diagnosis of acute symptomatic venous thromboembolism (proximal deep vein thrombosis or pulmonary embolism) objectively confirmed by previously described criteria (Wells et al., 2001, 2003). Patients were referred from the emergency room, outpatient clinics or directly from the radiologist or family physician after diagnosis. All patients were at least 18 years old. Exclusion criteria were: previous venous thromboembolic events, therapy with low molecular weight heparin for >48 h prior to testing, abnormal baseline international normalized ratio or activated partial thromboplastin time, active malignancy, occurrence of the venous thromboembolic event within 12 weeks of a transient risk factor (surgery, immobilization and hospitalization), pregnancy, presentation in the first 6 weeks postpartum, known chronic liver disease (defined as persistent elevation above normal levels of alanine aminotransfearase or gammaglutamyl transferase) and oral anticoagulant therapy (warfarin) in the previous 2 weeks. Thrombophilia screen A thrombophilia screen was drawn prior to the initiation of oral anticoagulant therapy and within 24 h of diagnosis of venous thromboembolism for all patients. Blood samples were collected in 0.109 m (3.8%) sodium citrate tubes (Becton-Dickinson, NJ, USA). In London, protein C was measured on an ACL 9000 Coagulometer (Instrumentation laboratory, Lexington, MD, USA) by the method of IL Test ProClot (Instrumentation Laboratory) and protein S was measured on an ACL 3000 Coagulometer by the method of IL-Test protein S (Instrumentation Laboratory). Samples were frozen and tested in batches on a weekly basis. The IL Test ProClot kit is a functional clotting protein C assay, based on the prolongation of an activated partial thromboplastin time (aptt) in the presence of activated protein C. Patient plasma is incubated with protein C deficient plasma, aptt reagent and protein C activator for a standard time. Coagulation is triggered by the addition of calcium chloride. The prolongation is directly proportional to the amount of protein C in the test sample. The IL Test protein S kit is a functional clot-based assay of free protein S. Patient plasma is incubated with activated protein S deficient plasma (protein C activator and protein S deficient plasma) and calcium bovine thromboplastin. The time to clot formation is measured. The protein S activity is proportional to the prolongation of the prothrombin time of a protein S deficient plasma to which diluted sample has been added. In Ottawa, protein C and protein S were measured on a STA (Diagnostica Stago, Mississauga, ON, Canada) prior to August 2002 and then on a BCS (Dade Behring Holdings, Deerfield, IL, USA). Prior to August 2002, protein C was measured by the method of Staclot PC (Diagnostica Stago) and then by the method of protein C Reagent (Dade Behring Holdings). Protein S was only measured by the method of ActiClot protein S (American Diagnostica, Montreal, Quebec, Canada), which assesses free protein S. Staclot Protein C measures functional Protein C based on the prolongation of aptt. Patient plasma is incubated with protein C deficient plasma and a protein C activator and then calcium chloride is added to begin measuring the aptt. The prolongation of the aptt is directly proportional to Protein C activity. The DadeBehring Protein C reagent method is similar. Patient plasma is incubated with protein C deficient plasma and a protein C activator. After incubation, an aptt is performed. The prolongation of the aptt is directly proportional to Protein C activity. Acticlot Protein S is a functional clot-based assay for free Protein S. Patient plasma is incubated with a reagent that contains factor Xa, activated Protein C and phospholipids, and then calcium chloride is added to initiate clot formation. The prolongation of the clotting time is directly proportional to the concentration of Protein S in the patient plasma. If a patient s initial testing for protein C and protein S was normal, no further testing was done and the patient was determined to have normal results for protein C and

M. J. Kovacs et al. 11 protein S. Any abnormal tests were repeated after at least 3 months of therapy provided the patient had discontinued warfarin for at least 14 days. For the repeat testing, if the attending physician was concerned about recurrent venous thromboembolism because of stopping warfarin, bridging therapy with dalteparin 5000 U S/C daily for 14 days was administered until warfarin could be restarted and a therapeutic International Normalized Ratio achieved. We determined the proportion of initial tests that were normal for protein C and protein S. We also calculated the proportion of abnormal tests that were confirmed on repeat testing. The 95% confidence intervals on these proportions were calculated according to the binomial distribution. Results There were 255 patients consisting of 125 men and 130 women with a mean age of 46 (range 18 86). One hundred and seventy-three patients had a diagnosis of deep vein thrombosis, 68 pulmonary embolism and 14 had concomitant deep vein thrombosis and pulmonary embolism. All patients had idiopathic (unprovoked) venous thromboembolism. Of the 255 patients, 254 were tested for protein C, 254 for protein S, and 253 were tested for both. The reason that two patients were not tested for protein S and C was administrative error. Of the 253 patients tested for both, 229 patients (91%; 95% confidence interval 87 94%) were negative for both protein C and protein S at the initial time of diagnosis. There were a total of 24 patients who were abnormal for either protein C or protein S, (no patients were abnormal for both protein C and protein S). Therefore, of 508 total initial tests (i.e. all protein C and protein S tests performed), 484 (95%; 95% confidence interval 93 97%) were normal. Twenty-one of the 24 abnormal results were retested off anticoagulants. Three patients were no longer in our practice (as they had Table 2. Abnormal protein S results (normal range: 0.60 1.60 U/ml) Baseline (U/ml) 0.23 0.60 0.44 0.77 0.23 0.22 0.43 0.46 0.47 0.44 0.51 1.25 0.33 0.20 0.53 0.37 0.51 0.81 0.51 1.15 0.46 0.45 Repeat (U/ml) Table 3. Abnormal protein C results (normal range: 0.72 1.23 U/ml) Baseline (U/ml) 0.33 0.75 0.55 0.78 0.40 0.95 0.36 0.36 0.44 0.49 0.65 0.76 0.54 0.83 0.33 1.07 0.68 0.65 0.60 0.65 Repeat (U/ml) moved from our region) and could not be retested. Ten (48%; 95% confidence interval 26 70%) were confirmed to be abnormal on retesting. Therefore, the false positive rate for all protein C and protein S tests performed in this study was 2.2% (95% confidence interval 1.1 3.9%) with 4.3% of patients having falsely abnormal results corresponding to a specificity of 96%. The results are summarized in Tables 1 3. Discussion Table 1. Results of protein C and protein S testing at presentation and at repeat testing if initial result was abnormal Protein C (n ¼ 254) Protein S (n ¼ 254) Normal 242 242 484 Abnormal 12 12 24 Repeat abnormal 4 6 10* Protein C and S* (n ¼ 508) *Three patients were lost to follow-up, hence, 10 of 21 patients had abnormal results on repeat testing. Thrombophilia screens are frequently performed in cases of idiopathic venous thromboembolism to determine the etiology, and/or the optimal duration of oral anticoagulant therapy. Clearly, Protein C and Protein S cannot be reliably measured while patients receive oral anticoagulant therapy with vitamin K antagonists. Moreover, it is widely believed that clotting-based thrombophilia screens (such as protein C and protein S) should not be performed at the time of diagnosis of acute venous thromboembolism because of a high false positive rate (D Angelo et al., 1988;

12 Protein C and protein S can be accurately determined Lane et al., 1996a, b; Cumming & Shiach, 1999; Kearon et al., 2000; Bauer, 2001; Seligsohn & Lubetsky, 2001; Thomas, 2001). Our study demonstrates that when protein C and protein S were assessed at the time of diagnosis of venous thromboembolism, the results were normal in 91% of patients, and 95% of the tests performed were normal enabling the physician to rule out these thrombophilias as contributing to the thromboembolic event. Nine percent of patients will require a repeat test. Although we did not confirm the normal results (see below) our study suggests that 98% of the initial results will be correct but positives must be confirmed with repeat testing off anticoagulants to detect the 2.2% of false positive results. This suggests that it is worthwhile to test for protein C and protein S at the time of initial diagnosis if thrombophilia testing is indicated. It is ordinarily recommended that even if abnormal results are obtained by testing first performed at a later date that these need to be confirmed by retesting anyway (Crowther & Kelton, 2003). D Angelo et al. (1988) were the first to suggest that protein S levels will be falsely decreased during acute deep vein thrombosis secondary to higher levels of C4bBP. This publication may have been the trigger for the perception that protein C and/or protein S levels are falsely lowered presumably due to consumption and that measurements should be performed at a later date. However, D Angelo s study reported on only eight patients and it simply describes lower mean protein S levels and does not indicate what proportion of results were below normal. The authors cautioned that the results needed to be confirmed in a larger study. To our knowledge ours is the first prospective study in a large series of patients to address this question. We did not perform repeat testing on patients with normal results. This is justified since the sensitivity of free protein S levels is 97.7% (Makris et al., 2000) and the sensitivity of protein C testing is likely similar. Application of Bayes Theorem is patients with low probability of disease (expected prevalence of protein S/C deficiency is at best 10% in this patient population) suggests the probability of the result being a false normal is 0.25% and thus repeat testing is not worthwhile (Bayes, 1991). Moreover, our overall rate of abnormal protein C and protein S results of 3.9% (10 of the 255 patients that were retested) is consistent with previous publications (Lane et al., 1996a, b; Cumming & Shiach, 1999; Kearon et al., 2000; Bauer, 2001; Seligsohn & Lubetsky, 2001; Thomas, 2001). Our results demonstrate that testing for protein C and protein S at the time of presentation of an acute venous thromboembolic event is useful to rule out deficiency of these proteins in the majority of patients. The false positive rate for tests in the study was only 2.2% (95% confidence interval 1.1 3.9%) with about half of the initially abnormal results confirmed on repeat testing. False positives will still occur likely as a consequence of the within-lab coefficient of variation described for these assays. The long-term within-lab analytical coefficient of variation is in the order of 8.6% for protein C and 14.1% for protein S and the coefficient of variation is even higher if the result obtained is blow normal (Meijer et al., 2003). Other reasons for the initial false positives include intraperson variation and regression to the mean (Fletcher, Fletcher & Wagner, 1998). Accurate screening for thrombophilia at the time of diagnosis is more convenient for patient care. Decisions about the optimal duration of oral anticoagulant therapy can be made much sooner if accurate results are determined at the initiation of therapy. Furthermore, family testing and counseling can be initiated earlier, if it is felt to be indicated (Green, 2001; Mannucci, 2001). Although we did not perform a formal cost-effectiveness analysis, as very few false positive tests occurred, it is probably more cost-effective to test patients at presentation than to later provide bridging with low molecular weight heparin while oral anticoagulants are stopped, in order to ascertain the protein C and protein S levels. This would be the case even if no false positives occurred during bridging therapy, an unlikely event. Finally, we only did one repeat to confirm positive test results and did not confirm positive test results by genetic analysis, which may be considered as a limitation of our study. Conclusions We conclude that if protein C and protein S are measured at the time of diagnosis of acute venous thromboembolism (prior to the initiation of oral anticoagulant therapy), the majority of the results (95%) will be normal and 98% of results will be accurate. Those that are abnormal need to be repeated as about half will be normal on repeat testing. It would seem reasonable to suggest this as a routine practice with the proviso that any abnormal results be confirmed at a later date. Acknowledgements This manuscript contains original data. All authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Dr Kovacs is an Internal Scholar of the

M. J. Kovacs et al. 13 Department of Medicine at the University of Western Ontario. Dr Rodger is the recipient of the Maureen Andrew New Investigator Award from the Heart and Stroke Foundation of Ontario. Dr Wells is a Canada Research Chair in Thromboembolic Diseases. References Bauer K.A. (2001) The thrombophilias: well-defined risk factors with uncertain therapeutic implications. Annals of Internal Medicine 135, 367 373. Bayes T. (1991) An essay towards solving a problem in the doctrine of chances. 1763. MD Computing 8, 157 171. Crowther M.A. & Kelton J.G. (2003) Congenital thrombophilic states associated with venous thrombosis: a qualitative overview and proposed classification system. Annals of Internal Medicine 138, 128 134. Cumming A.M. & Shiach C.R. (1999) The investigation and management of inherited thrombophilia. Clinical and Laboratory Haematology 21, 77 92. D Angelo A., Vigano-D Angelo S., Esmon C.T. & Comp P.C. (1988) Acquired deficiencies of protein S. Protein S activity during oral anticoagulation, in liver disease, and in disseminated intravascular coagulation. Journal of Clinical Investigation 81, 1445 1454. Fletcher R.H., Fletcher S.W. & Wagner E.H. (1998) Abnormality. In: Clinical Epidemiology: The Essentials (ed. N. Collins), pp. 19 41. Williams & Wilkins, Baltimore, MD. Green D. (2001) Genetic hypercoagulability: screening should be an informed choice. Blood 98, 20. Kearon C., Crowther M. & Hirsh J. (2000) Management of patients with hereditary hypercoagulable disorders. Annual Review of Medicine 51, 169 185. Lane D.A., Mannucci P.M., Bauer K.A., Bertina R.M., Bochkov N.P., Boulyjenkov V., Chandy M., Dahlback B., Ginter E.K., Miletich J.P., Rosendaal F.R. & Seligsohn U. (1996a) Inherited thrombophilia: part 1. Thrombosis and Haemostasis 76, 651 662. Lane D.A., Mannucci P.M., Bauer K.A., Bertina R.M., Bochkov N.P., Boulyjenkov V., Chandy M., Dahlback B., Ginter E.K., Miletich J.P., Rosendaal F.R. & Seligsohn U. (1996b) Inherited thrombophilia: part 2. Thrombosis and Haemostasis 76, 824 834. Makris M., Leach M., Beauchamp N.J., Daly M.E., Cooper P.C., Hampton K.K., Bayliss P., Peake I.R., Miller G.J. & Preston F.E. (2000) Genetic analysis, phenotypic diagnosis, and risk of venous thrombosis in families with inherited deficiencies of protein S. Blood 95, 1935 1941. Mannucci P.M. (2001) Genetic hypercoagulability: prevention suggests testing family members. Blood 98, 21 22. Meijer P., Kluft C., Haverkate F. & de Maat M.P. (2003) The longterm within- and between-laboratory variability for assay of antithrombin, and proteins C and S: results derived from the external quality assessment program for thrombophilia screening of the ECAT foundation. Journal of Thrombosis and Haemostasis 1, 748 753. Salzman E.W. & Hirsh J. (1994) The epidemiology, pathogenesis and natural history of venous thrombosis. In: Hemostasis and Thrombosis: Basic Principles and Clinical Practice (eds R.W. Colman, J. Hirsh, V.J. Marder & E.W. Salzman), 3rd edn, pp. 1275 1296. JB Lippincott, Philadelphia. Seligsohn U. & Lubetsky A. (2001) Genetic susceptibility to venous thrombosis. New England Journal of Medicine 344, 1222 1231. Silverstein M.D., Heit J.A., Mohr D.N., Petterson T.M., O Fallon W.M. & Melton L.J. III (1998) Trends in the incidence of deep vein thrombosis and pulmonary embolism: a 25-year population-based study. Archives of Internal Medicine 158, 585 593. Thomas R.H. (2001) Hypercoagulability syndromes. Archives of Internal Medicine 161, 2433 2439. Weinmann E.E. & Salzman E.W. (1994) Deep-vein thrombosis. New England Journal of Medicine 331, 1630 1641. Wells P.S., Anderson D.R., Rodger M., Stiell I., Dreyer J.F., Barnes D., Forgie M., Kovacs G., Ward J. & Kovacs M.J. (2001) Excluding pulmonary embolism at the bedside without diagnostic imaging: management of patients with suspected pulmonary embolism presenting to the emergency department by using a simple clinical model and d-dimer. Annals of Internal Medicine 135, 98 107. Wells P.S., Anderson D.R., Rodger M., Forgie M., Kearon C., Dreyer J., Kovacs G., Mitchell M., Lewandowski B. & Kovacs M.J. (2003) Evaluation of d-dimer in the diagnosis of suspected deep-vein thrombosis. New England Journal of Medicine 349, 1227 1235.