Prolonged Prothrombin Time After Discontinuing Vitamin K Antagonist

A 65-year-old man was admitted to our hospital for a colonoscopy. He had been taking a vitamin K antagonist (acenocoumarol) since 1989, when he suffered from 2 episodes of lower limb deep vein thrombosis. Over time, his prothrombin time (PT)4 international normalized ratio (INR) values were stable and checked regularly at an anticoagulation clinic. The anticoagulation clinic advised him to stop the intake of vitamin K antagonists 3 days before the colonoscopy. For patients undergoing a colonoscopy, a maximum INR of 1.7 on the day of the procedure is accepted. At our hospital, on the day of the scheduled colonoscopy procedure, our patient had an INR of 2.5. Based on this value, the colonoscopy was canceled and rescheduled. In addition to the INR, an activated partial thromboplastin time (APTT) was performed, which was greatly prolonged (>125 s). The patient accepted the cancellation of the procedure but questioned his INR result of 2.5. Therefore, a few hours after visiting our hospital, he had the values checked again at his anticoagulation clinic, which resulted in an INR of 1.2. Puzzled, he phoned the clinical laboratory of the hospital and asked to speak with the clinical chemist for an explanation of these inconsistent results. Searching for the cause of the discrepant INR results, the clinical chemist initially considered preanalytical errors (1). Reviewing the tube, it was appropriately filled and the hematocrit of the patient was normal, excluding the possibility of an incorrect citrate to plasma ratio. Underfilling of the tube is probably the most important preanalytical error. The recommended ratio of blood to sodium citrate anticoagulant in tubes is 9:1. Underfilled tubes contain proportionally more sodium citrate per volume of plasma, binding more calcium and potentially leading to longer clotting times (2). Other prerequisites for optimal sample integrity, such as processing within 24 h for PT-INR and 4 h for APTT, were met (3). Blood was drawn by venipuncture for coagulation testing only, excluding the possibility of heparin carry-over from a line or cross-contamination by tube additives. Although most PT-INR reagents use heparin neutralization, cases showing prolonged INR resulting from heparin sensitivity have been described (4). What are possible causes of a PT-INR that differs between laboratories? Considering the medical history of the patient, what diagnosis does the prolonged APTT suggest? What additional testing should be done to confirm the diagnosis in this patient? After reviewing the medical history of the patient and learning about the thrombotic events that occurred at the age of 39 years, the clinical chemist suggested testing for antiphospholipid syndrome (APS). A week later, blood was drawn for lupus anticoagulant (LA), anticardiolipin antibodies, and anti-β2-glycoprotein antibodies tests. These tests were performed while the patient was using his vitamin K antagonist, as he restarted this medication on the day of the canceled colonoscopy. The tests for APS, INR, and APTT were performed by 2 independent laboratories, the results of which are shown in Table 1. Both laboratories found a strong presence of LA. Because our patient was using vitamin K antagonists, the patient's plasma was diluted with normal plasma to reduce the effect of medication before performing LA coagulation assays. In addition, both anticardiolipin IgG and anti-β2-glycoprotein IgG antibodies were found, whereas IgM antibodies were not detectable. Interestingly, large differences were seen in the reported INR and APTT values between the 2 laboratories. Results of coagulation testing in two different hospital laboratories. INR reagent, Innovin (Siemens); APTT reagent, Actin-FSL (Siemens). INR reagent, Hepato-Prest (Stago); APTT reagent, NM. INR reagent, Thromborel S (Siemens); APTT reagent, Actin-FS (Siemens). NM, not measured. Results of coagulation testing in two different hospital laboratories. INR reagent, Innovin (Siemens); APTT reagent, Actin-FSL (Siemens). INR reagent, Hepato-Prest (Stago); APTT reagent, NM. INR reagent, Thromborel S (Siemens); APTT reagent, Actin-FS (Siemens). NM, not measured. APS is an autoimmune disorder characterized by recurrent venous or arterial thromboses and pregnancy loss. Consensus criteria for the diagnosis of APS are shown in Table 2 and include 3 laboratory tests: LA, anticardiolipin antibodies, and β2-glycoprotein antibodies. At least 1 of these tests needs to be (persistently) positive to consider the diagnosis of APS (5, 6). Triple-positivity for the 3 tests, as in our patient, is strongly associated with thromboembolic events (7). Classification criteria for the antiphospholipid syndrome.a From Keeling et al. (5), as adapted from Miyakis et al. (6). Reproduced with permission. Classification criteria for the antiphospholipid syndrome.a From Keeling et al. (5), as adapted from Miyakis et al. (6). Reproduced with permission. Whereas the APTT is usually prolonged in those with APS, it is a rare phenomenon for LA to cause a prolonged PT. In the 1990s, several reports raised concerns about the use of the INR for monitoring oral anticoagulation in patients with APS [summarized by Della Valla, (8)], showing that assays using recombinant thromboplastin are particularly sensitive to LA interference. However, in 2001, Tripodi et al. compared the influence of LA on INR using 9 different commercial thromboplastins in 58 cases of APS and concluded that the interference of LA on INR was not sufficient to cause major concerns with the majority of commercial thromboplastins (9). Later, Ferrazzi et al. compared 2 commercial thromboplastins, 1 conventional and 1 recombinant, and found no difference in the effect of LA presence on INR values <4 (10). Moore et al. tested 400 patients with LA who were not receiving therapeutic anticoagulation and found a prolonged PT with a recombinant thromboplastin in 4.3% of patients (11). Recently, Isert et al. reported that INR variability did not differ with the use of 4 different reagents in patients taking vitamin K antagonists regardless of any history of APS (12). In addition, they investigated the agreement between conventional INR values and INR values derived from a chromogenic factor X assay, which has been suggested as an alternative to monitor vitamin K antagonist therapy in patients suspected of falsely increased INR values. In our case, the highest INR was found in the laboratory using a reagent containing recombinant human tissue thromboplastin (Innovin from Siemens). A recent communication from the Scientific and Standardization Committee of the International Society on Thrombosis and Hemostasis gives useful recommendations regarding the monitoring of APS patients treated with vitamin K antagonists (13). APS is a common cause of a prolonged APTT, but high titers of antiphospholipid antibodies may occasionally lead to a prolonged PT as well. The effect of antiphospholipid antibodies on APTT and PT may vary greatly between reagents from different manufacturers. Testing for antiphospholipid antibodies is indicated for any unexpected result in coagulation testing, when preanalytical causes have been excluded and results are not corrected in mixing studies. It is advisable to check the INR of the patient after discontinuing vitamin K antagonists in the same anticoagulation clinic where the patient is regularly seen. This case shows that a strong presence of antiphospholipid antibodies can result in large heterogeneity of INR values, and may delay treatment and thereby increase healthcare costs. The influence of antiphospholipid antibodies on INR values is dependent on the reagent used by the clinical laboratory. Therefore, doubtful results in coagulation testing should always be discussed with a laboratory professional, as was done in this case by the patient himself. Colonoscopy was successfully performed after the patient stopped anticoagulant therapy for 3 days. The patient continued using a vitamin K antagonist for his APS. prothrombin time international normalized ratio activated partial thromboplastin time antiphospholipid syndrome lupus anticoagulant.