Controversies in heparin monitoring

A 57-year-old man with end-stage liver disease is hospitalized for liver transplant evaluation. He has a history of portal hypertension and esophageal varices, but no bleeding history. His baseline PT is 17 sec (normal 11–13 sec) and aPTT is 46 sec (normal 21–34 sec). While in hospital, the patient developed a femoral vein DVT. The patient was started on unfractionated heparin (UFH) using a nomogram targeting an anti-Xa range of 0.3–0.7 U/ml. The following day the anti-Xa level was 0.7, the upper end of the therapeutic range, and the aPTT was >300 sec. A hematology consult was obtained to address the following questions: Why are the tests discordant? Is this patient at increased risk of bleeding? Should he be anticoagulated, and if so, what range is appropriate, and which test should be used? We will try to answer these questions after we consider the many issues involved with therapeutic monitoring of UFH. Heparin was discovered in 1916 by a medical student, Jay MacLean, working in the laboratory of William Howell at Johns Hopkins University [1]. Ironically, Howell was screening dog liver extracts for procoagulant activity, but instead found anticoagulant activity. Howell named the compound heparin after the hepatic source. Heparin was localized to secretory granules in mast cells by Hjalmar Holtgren in 1936. Erik Jorpes demonstrated that heparin was a sulfated glycosaminoglycan [2]. Mucosal mast cells remain the source of all commercial forms of heparin. The physiology of mast cell heparin is incompletely understood [3]. Heparin is required for the production of mast cell proteases [4, 5]. It is believed that heparin would be present at high concentrations locally in setting of mast cell degranulation but the role of heparin in terms of local anticoagulation in this circumstance is unknown. Heparin contains a unique highly sulfated pentasaccharide sequence which is required for anticoagulant effect. This anticoagulant effect is mediated by binding to the serine protease inhibitor antithrombin (also known as serpin C1). Heparin binding induces an allosteric change in antithrombin with resultant greater catalytic activity [6]. Heparin also serves to colocalize thrombin, creating a ternary enzyme-substrate cofactor complex, and facilitate this specific serpin inhibition of thrombin. The mechanism of antithrombin inhibition of thrombin has been recently reviewed [7] and involves presentation of an antithrombin reactive site loop which following nucleophilic attack by thrombin at the R393-S394 bond, results in a stable enzyme-inhibitor complex which is thus unavailable to inactivate serine proteases in the clotting cascade. Heparin was found to be present in many tissues, including lung and intestine. As large scale preparation from the original canine source was not feasible, Charles and Scott working at the University of Toronto purified heparin from ox lung [8]. Following the description of purification and wider availability, heparin began to be used for prevention and treatment of thrombosis. Before the advent of heparin therapy, treatment of phlebitis consisted of bed rest and observation, with a significant proportion of individuals extending clot and suffering fatal pulmonary embolism. A 1959 review of heparin therapy by Bauer noted that in the era before heparin therapy, in a series of 29 patients with diagnosis of lower extremity DVT by venography, in 24/29 cases the process spread to the femoral vein within days, in 10/ 29 cases the other leg developed DVT, and in 11/29 cases pulmonary emboli ensued, 2 of which were fatal. In the 1940s clinical trials of heparin anticoagulation had started worldwide [9]. Bauer's group began treating DVT with intermittent intravenous injections of heparin ranging from 100 to 150 mg 3–4 times/day, with early ambulation. This therapy appeared to dramatically alter the course of venous thrombophlebitis. A follow-up series of 627 cases treated with heparin reported in 1954 by the same group reported a mortality of 0.8%, recurrence in 2.7% and bleeding in 2% [9]. No monitoring of heparin was used or felt to be necessary given the low incidence of adverse events observed. The first randomized controlled trial of heparin therapy was reported by Barritt and Jordan from Bristol, United Kingdom in 1960 [10]. Patients were diagnosed with VTE by clinical signs and symptoms and randomized to observation or therapy with UFH and nicoumalone (a coumarin vitamin K antagonist). Heparin was given as 10,000U IV every 6 hrs for 6 doses without coagulation monitoring, and the coumarin adjusted to PT 2-3X control for a total of 14 days of therapy. An interim analysis of the first 35 patients showed 5/16 died from PE in the observation arm, versus 0/16 in the treatment arm, a clinically dramatic and statistically significant result. Thirty-eight subsequent patients were enrolled in the treatment arm, with no further fatal pulmonary emboli. This study strongly influenced clinical practice in favor of treating PE, and by analogy DVT, with heparin and coumarins [11]. The commonly used aPTT therapeutic range of 1.5–2.5 times control levels for the treatment of venous thromboembolic disease originated from a report of an observational prospective clinical trial of heparin therapy reported by Basu et al from McMaster University in 1972 [12]. Two hundred and thirty-four patients were studied, of which 162 had VTE based on clinical criteria, supported by CXR and lung scan in 81%. Patients were treated with a 5000U bolus of IV heparin, followed by 24,000U/day, with a recommendation to titrate to an aPTT of 1.5–2.5X control. Patients with recurrent VTE were given similar amount of heparin compared to those without recurrent thrombosis, but the aPTT was significantly lower in the recurrent thrombosis group. The bleeding rate was 8% without significant differences in heparin dose or aPTT, and higher in surgical patients than medical patients. Based on these data, the authors recommended keeping an aPTT > 1.5X control to reduce risk of VTE recurrence. The range of 1.5–2.5X control was correlated with a therapeutic range of 0.2–0.4 U/ml by protamine in a rabbit model of venous thrombosis [13]. The anti-Xa range corresponding to a protamine titration range of 0.2–0.4 U/ml was reported to be 0.35–0.67 U/ml at the same institution [14]. A subgroup analysis of 3 randomized trials reported that the risk of recurrent VTE was significantly higher in patients who failed to achieve a therapeutic aPTT in the first 24 hrs of therapy [15]. However, subsequent studies suggested that as long as UFH was administered as an initial bolus followed by a total dose of at least 30,000 U/24 hrs, the risk of recurrent thromboembolic disease was independent of aPTT result. In a meta-analysis of five trials in the treatment of venous thromboembolic disease, provided patients were treated with initial heparin bolus and at least 30,000 U/24 hrs, the risk of recurrent venous thromboembolic events were similar between those whose aPTT results were sub therapeutic for the first 24–48 hrs (6.3%) and those whose aPTT results where greater than the lower limit of the therapeutic range (7%) [16]. A subsequent analysis of three trials in the treatment of acute VTE treated with an initial bolus of UFH followed by a minimum of 30,000 U/24 hrs found a slight but nonsignificant increase in recurrent VTE in those with sub therapeutic vs. therapeutic aPTT results in the first 24 hrs [17]. Thus, the utility of monitoring heparin therapy in the treatment of acute thromboembolic disease remains controversial. UFH is also monitored to prevent bleeding, under the assumption that the higher the measured "anticoagulant" effect, the higher the risk of major bleeding complications. An increase in major bleeding has been seen with higher heparin doses [18] or supra-therapeutic aPTT results in some, but not all, studies [19, 20]. Hull et al. [21]randomized patients with proximal venous thromboembolic disease to either 5 or 10 days of UFH followed by warfarin. In this study, major bleeding was not increased in those with supratherapeutic aPTT results (3.2%) compared to those without supratherapeutic aPTT results (9.4%) [15]. In contrast, Anand et al performed a subgroup analysis of the OASIS-II study, a phase III trial of patients with non-ST elevation acute coronary syndromes randomized to receive UFH or hirudin. In this analysis, for every 10 sec increase in the aPTT, the probability of major bleeding was increased by 7% [21]. Patient-specific factors are perhaps the most important contributor to bleeding risk in patients on heparin therapy. Factors that increase the risk for bleeding include recent procedures or trauma, older age, comorbid disease, and other hemostatic defects [20]. UFH, low molecular weight heparins (LMWH) and fondaparinux have been shown to be of comparable anticoagulant efficacy in many clinical trials. However, while UFH will significantly prolong the aPTT at therapeutic concentrations, LMWH and fondaparinux have minimal effects on the aPTT. This difference in effect on the aPTT may be due to the greater inhibition of thrombin relative to Xa by UFH as discussed above, whereas LMWHs and fondaparinux have more anti-Xa activity. However, this observation raises the question: What is the aPTT actually measuring? If the aPTT specifically measured therapeutic anticoagulant effect, then all of the heparin family members would be expected to give similar results in the range that is associated with clinical benefit. Many patient variables interfere with use of the aPTT to assess heparin effect. Patients with lupus anticoagulants may have very prolonged aPTT, but this is not associated with bleeding risk. The use of warfarin results in variably increased sensitivity of the aPTT to heparin, depending on the assay reagents. Patients with inherited or acquired factor deficiencies, such as those with liver disease, may also have an increased baseline aPTT. In the case of factor deficiency, the impact on hemostasis will depend on which factor is deficient. For example, a patient with factor XII deficiency may have a prolonged aPTT, however will not have an increased risk of bleeding, whereas a patient with factor VIII deficiency and a similarly prolonged aPTT will have a high risk of bleeding. Patients with acute inflammatory states may also manifest heparin resistance [14, 22, 23]. Perhaps the biggest problem associated with using the PTT to measure heparin effect, independent of patient variables, is the lack of standardization of methods. Thus, different reagent-instrument combinations may lead to significant differences in the aPTT time in seconds for a given heparin concentration [24-26] This can be appreciated by looking at any College of American Pathology heparin survey, where the aPTT value in seconds is widely variable depending on the reagent/instrument combination used. A recent study examined 15 different aPTT reagents on the same instrument platform and found that the aPTT ratios corresponding to an anti-Xa level from 0.3 to 0.7 ranged from 1.6 to 2.6 with one reagent to an extreme of 3.6–12. To address this issue, consensus panels have recommended that laboratories calibrate their aPTT assays by measuring the anti-Xa level, based on the ability of heparin to neutralize fixed amounts of factor Xa added to plasma, or by protamine titration [27-29] However, the validity of this approach has recently been questioned, as such calibration methods may not enhance interlaboratory agreement in UFH monitoring [16, 30]. The main arguments in favor of aPTT are clinician familiarity and comfort; from a laboratory standpoint the tests are widely available, inexpensive, easy to perform, and have high precision for any given method, with low coefficients of variation and excellent interlaboratory agreement. The aPTT also offers a more global assessment of the integrity of the hemostatic system. The main problems with the aPTT are the lack of standardization for use in heparin monitoring and the many potential test interferences noted above. Alternatively the anti-Xa chromogenic assay can be used to directly measure heparin effect [31] This method has the advantage of less interference from patient variables mentioned above. These methods can now be run on automated instruments in the same time frame as aPTT assays. However, standardization is also an issue with anti-Xa assays.[26, 32, 33] In addition, the precision of anti-Xa assays is lower than for PTT assays [34]. Outside of coagulation specialists there is a lack of familiarity and clinical comfort with using anti-Xa assays for UFH monitoring. Anti-Xa measurements do not give an overall assessment of hemostasis, are not routinely available in many hospital laboratories and are more costly. There is less outcome data for anti-Xa methods. In general, there is a lack of data correlating these methods of measuring heparin effect in patients and outcome data validating any either approach as ideal. To address this question, we recently evaluated a large cohort of hospitalized adults on UFH with paired aPTT and anti Xa values (n = 2,321 paired values from 539 patients). All simultaneously drawn aPTT, PT and anti-Xa values over a 6 month period were obtained. Based on a polynomial fit, 3 groups were defined: patients who consistently had higher aPTT values than expected for a given anti-Xa (PTT > anti-Xa); patients whose PTT values consistently fell within the expected range for a given anti-Xa activity (aPTT = anti-Xa); and patients that had at least two consecutive aPTT > anti-Xa. Electronic medical records were reviewed for outcomes of bleeding, thrombosis and mortality within 30 days. We found that data pairs were frequently discordant. Nine hundred and seventy-one (42%) were in the aPTT > anti-Xa group (i.e., the PTT was disproportionately long compared to the anti-Xa result). The aPTT > anti-Xa group was strongly associated with an INR ≥ 1.5 (P < 0.0001). Many of these patients were on concomitant warfarin therapy. Patients with consistent aPTT > anti-Xa on heparin alone (not on concomitant warfarin) had the highest 30-day mortality (15 of 39, 38%), compared to a 30-day mortality of 6% (3 of 48) for aPTT > anti-Xa patients who were on both heparin and warfarin (P < 0.001) [35]. The cause of death was related to progression of underlying disease in most cases; there were no fatal bleeding events. However, patients with at least 2 consecutive aPTT > anti- Xa values had an increased risk of bleeding compared to the aPTT = anti-Xa group. Thus, in this large retrospective cohort analysis, patients with disproportionate prolongation of aPTT relative to anti-Xa activity did have worse outcomes and constituted a high risk group. If these data are prospectively confirmed, there may be utility in measuring both aPTT and anti-Xa to stratify risk and make dosing decisions. Laboratory monitoring of UFH therapy remains a controversial topic 96 years following the discovery of this drug. While the available evidence suggests that heparin could be given without monitoring for patients with VTE, it is unlikely that any institution would adopt such a policy. There are several reasons for this. Most uncomplicated VTE cases are now managed on an outpatient basis with LMWH or fondaparinux without monitoring. The current standard is to transition to warfarin with INR monitoring, but it is likely that in the near future with the use of the newer oral antithrombin and anti-Xa medications that many of these patients will be treated for VTE without monitoring of any kind. Use of UFH therefore is restricted to inpatient populations that tend to be sicker and have multiple comorbidities. The pharmacokinetics of UFH are complex and dose-dependent and these uncertainties can be magnified in patients with multisystem disease. In the inpatient setting monitoring may be beneficial to detect medication errors which could cause unintended overdosages and increased bleeding risk or underdosages leading to increased thrombosis risk. Heparin is also commonly used for other indications where the therapeutic target is less well defined than for VTE. Recognizing that errors involving heparin and other anticoagulants are frequent and associated with significant patient risk, in 2008 the Joint Commission added a requirement to the National Patient Safety Goals for institutions to define an anticoagulation management program for patients on warfarin, UFH and LMWHs. This program dose not specify a particular monitoring scheme, but mandates hospitals to standardize practice and reduce adverse drug events and likelihood for patient harm associated with anticoagulant therapy. Thus given the above considerations and the current regulatory environment, standardized infusion and monitoring schemes for UFH therapy are a given in the United States. From the forgoing discussion, neither the aPTT nor anti-Xa based monitoring is a perfect solution. The use of combined aPTT and Xa measurements may help inform treatment decisions and prognosis as illustrated in the case vignette. Let us now return to the case presentation—a man with end-stage liver disease, with prolonged baseline PT and aPTT who develops a DVT. Twenty-four hours following initiation of heparin per hospital protocol, the aPTT was >300 sec and the anti-Xa value was 0.7 U/ml, at the upper end of the suggested therapeutic range of 0.3–0.7. This patient had a prolonged PT, which is known to increase the sensitivity of the aPTT to heparin therapy. Additionally his baseline aPTT was prolonged. Mixing studies performed on the patient samples obtained before initiation of heparin therapy demonstrated correction, consistent with coagulation factor deficiency due to underlying liver disease. Thus in this individual, the aPTT cannot be used to reliably estimate heparin effect. A common misperception of patients with liver disease and prolonged baseline clotting times is that they are functionally anticoagulated and protected from VTE. These patients have defects in both pro-coagulant and anticoagulant systems and are more fragile than patients with normal liver function [36]. Thus they are at increased risk for both bleeding and thrombosis, and depending on clinical circumstances one or both may manifest. Obviously given his underlying liver disease with hepatic synthetic dysfunction and portal hypertension, this patient should be considered to be at increased bleeding risk. This case demonstrates common patient-specific interferences (baseline prolongation of clotting times) which make it difficult to use the aPTT to manage UFH therapy. The decision to anticoagulate or not depends on the clinical assessment of risk of extension of clot and pulmonary embolism versus risk of bleeding. One option might be that you decide the risk of anticoagulation is too high and elect for a vena caval filter to decrease the risk of pulmonary embolism. If the decision is made to continue anticoagulation with UFH, using an anti-Xa assay and targeting the lower end of the therapeutic range (e.g., 0.3–0.5 U/ml) may represent the best balance of the need for anticoagulation with risk of bleeding in this individual. Finally with regard to the question: which test is better? As discussed above, the aPTT or the anti-Xa assay both have their strengths and weaknesses. However the tests are to some extent complimentary and use of both assays may be informative in complex hospitalized patients. Our practice is to obtain baseline coagulation tests before initiation of anticoagulation therapy. If these are abnormal, then the reason for this is determined and the risk-benefit equation reconsidered. If anticoagulation with UFH is indicated, then heparin is given by a standardized nomogram in by electronic order entry. The first 3 assessments of heparin effect include both aPTT and Xa. If these are concordant, then subsequent heparin monitoring is performed with anti-Xa activity as there are less patient interferences with this measurement. If the tests are discordant as in the case above, then there is an opportunity to reassess risk and management strategy. Our large retrospective series suggests that patients with disproportionate prolongation of aPTT relative to anti-Xa values are at increased risk of bleeding and mortality. However, confirmation of these findings in a prospective cohort would be required to broadly recommend this practice.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.