Guidelines on the use and monitoring of heparin

The guideline group was selected to be representative of UK-based medical experts. The drafting group met and communicated by e-mail. Draft guidelines were revised by consensus. Since the initial guideline published by the British Committee for Standards in Haematology (BCSH; Colvin & Barrowcliffe, 1993) evidence-based guidelines on the use and monitoring of heparin have been included in the American College of Chest Physicians Consensus Conferences on Antithrombotic Therapy (ACCP; Hirsh & Raschke, 2004) and the Scottish Intercollegiate Guidelines Network (SIGN; http://www.sign.ac.uk/guidelines/fulltext/36/section12.html). Reference to these guidelines is advised for a comprehensive review of the evidence. The recommendations in this BCSH guideline generally reflect those of the ACCP and SIGN and are updated where appropriate to encompass recent studies. The guideline was reviewed by a multidisciplinary sounding board, the BCSH and the British Society for Haematology (BSH) and comments incorporated where appropriate. Criteria used to quote levels and grades of evidence are as in Appendix 3 of the Procedure for Guidelines Commissioned by the BCSH (http://www.bcshguidelines.com). The target audience for this guideline is healthcare professionals involved in the management of patients receiving heparin. This guideline will be reviewed in 2008. Interim addendums will be published as required on the BCSH website (http://www.bcshguidelines.com). Heparin remains the most widely used parenteral antithrombotic. The general adoption of low-molecular weight heparins (LMWHs) represents a significant therapeutic advance in terms of ease and convenience of administration. There may also be some advantages in terms of efficacy and fewer side-effects. A further development has been the introduction into clinical practice of the synthetic pentasaccharide factor Xa inhibitor, fondaparinux. This compound may have additional advantages although its role in prophylaxis and treatment has yet to be fully defined. Heparin is a naturally occurring glycosaminoglycan produced by the mast cells of most species. The pharmaceutical drug is extracted from porcine or bovine mucosa. All the products currently used in the United Kingdom are of porcine origin. Heparin consists of alternating chains of uronic acid and glucosamine, sulphated to varying degrees, and has a molecular weight (MW) range of 5000–35 000 Da. Samples of heparin over the last 50 years have shown a steady rise in MW with a concomitant rise in specific activity (Mulloy et al, 2000); current preparations have a mean MW of about 13 000–15 000 Da and specific activity of 180–220 IU/mg. Although unfractionated heparin (UFH) is still employed, for many indications there has been a trend towards use of fractionated or LMWHs. These are manufactured from UFH by controlled depolymerisation using chemical (nitrous acid or alkaline hydrolysis) or enzymatic (heparinase) methods. Although the processes yield different end groups, there is no evidence that these differences in chemical structure affect biological function. The biological properties of any LMWH are primarily determined by its MW distribution. As shown in Table I, the products currently available for clinical use have an average MW between 3000 and 5000 Da. They are heterogeneous in MW, although the polydispersity is less than that of UFH and 60–80% of the total polysaccharides lie between MW 2000 and 8000 Da. All anticoagulant properties of UFH and LMWH depend on the presence of a specific pentasaccharide sequence, which binds with high affinity to antithrombin and potentiates its activity (Lindahl et al, 1979). This sequence is present in about one-third of the chains in UFH but in lower proportions in LMWHs because some of these sequences are destroyed by the depolymerisation process. Acceleration of inhibition of factor Xa (anti-Xa activity) requires only the pentasaccharide sequence (approximate MW 1700 Da), but potentiation of thrombin inhibition [anti-IIa activity, also prolongation of activated partial thromboplastin time (APTT)] requires a minimum total chain length of 18 saccharides (MW approximately 5400 Da; Lane et al, 1984). Therefore, in all LMWH preparations the anti-Xa activity exceeds the anti-IIa activity. The ratio of anti-Xa to anti-IIa activity varies between 1·6 to 4·2 for all except one product, which has the lowest MW and an anti-Xa to anti-IIa ratio of 9·6 (Table I). There has been much debate about the relative importance of anti-Xa and anti-IIa activity in the anticoagulant effect of LMWHs in vivo. Biochemical studies have shown that the fractions with anti-IIa activity are largely responsible for the inhibition of thrombin generation in plasma (Béguin et al, 1988). However, fractions with only anti-Xa activity have antithrombotic activity in animals, including the synthetic pentasaccharide, which has now been shown to be an effective antithrombotic agent in clinical trials (Buller et al, 2003). Dosage of the various LMWHs correlates better with anti-Xa rather than anti-IIa activity (Barrowcliffe, 1995), and for practical purposes, anti-Xa activity is the only measurement that can be used for monitoring LMWHs. Overall, it seems likely that both types of action contribute to the antithrombotic effects of LMWHs, although the efficacy of fondaparinux as an antithrombotic indicates that anti-Xa activity alone is also effective. Unfractionated heparins are assayed in either International Units (IU), defined by the World Health Organization (WHO) International Standard, or United States Pharmacopoeial (USP) Units, defined by the USP standard and assay method. There is a 7–10% difference between the two units. Most products available in Europe are assayed in IU, using the European Pharmacopoeia method, based on prolongation of the APTT of sheep plasma. When initially developed, LMWHs were assayed against the UFH standard by a variety of different methods and the units for the different products could not be compared readily. An international standard for LMWHs was established in 1986 (Barrowcliffe et al, 1988). Unfractionated heparin is available as sodium or calcium salt. After subcutaneous injection of equal amounts the overall anticoagulant activity is lower with calcium than with sodium salt, but this does not affect clinical efficacy. There may be a lower incidence of ecchymoses after subcutaneous injection of the calcium than of the sodium salt, but there is no clear evidence for any major differences in the incidence of other haemorrhagic effects. All LMWHs are in the sodium form except fraxiparine (Nadroparin), which is a calcium salt. Both UFH and LMWHs are given parenterally, either by intravenous or subcutaneous injection. Metabolism is by a saturable mechanism, involving binding to endothelial cells and clearance by the reticuloendothelial system, and a non-saturable mechanism involving mainly renal clearance. Both mechanisms are important for UFH, but renal clearance predominates for LMWHs. This is clinically important as accumulation of LMWHs may occur, with increased bleeding risk, in renal failure. There is no evidence that UFH or LMWHs cross the placenta. The principal way in which UFH and LMWHs differ is in their pharmacokinetics. Low-molecular weight heparins have longer half-lives than UFH, after both intravenous and subcutaneous injection (Boneu et al, 1990). The intravenous half-life is about 2 h, measured as anti-Xa activity, although somewhat shorter (about 80 min) when measured by anti-IIa assay. The half-life of UFH is dose-dependent but, at usual intravenous doses, it is 45–60 min by both assay methods. The subcutaneous half-life of LMWHs is about 4 h, measured as anti-Xa activity although there are some differences in pharmacokinetic profiles between LMWHs. Unlike subcutaneous UFH, which has a bioavailability of <50%, all LMWHs have a bioavailability after subcutaneous injection of 90–100%. These differences in pharmacokinetics and bioavailability are responsible for the successful clinical use of once daily subcutaneous injections of LMWH. Several proteins interact strongly with heparin to antagonise its anticoagulant activity, the most important being platelet factor 4 (PF4) and protamine. Binding affinity to these proteins is reduced with decreasing MW, so that LMWH preparations require higher concentrations of PF4 or protamine to neutralise their activity than does UFH. Below 18 saccharides, heparin chains become increasingly resistant to neutralisation by either of these agents, so that all LMWHs have a portion of their anti-Xa activity that is non-neutralisable (Holmer & Söderström, 1983; Lane et al, 1984). Animal studies suggest that this does not affect the ability of protamine to neutralise the haemorrhagic action of LMWHs (Diness & Ostergaard, 1986), although the lack of a fully efficient method of reversal of LMWHs has been raised as a concern in relation to clinical practice (see below). Low-molecular weight heparins bind less strongly than UFH to endothelial cells, and this is partly responsible for the difference in pharmacokinetics, because endothelial binding and processing is an important mechanism of clearance for UFH. LMWHs also interact with platelets less readily than UFH, whether measured as potentiation of spontaneous aggregation or inhibition of agonist-induced aggregation (Salzman et al, 1980). Finally, LMWHs release lower concentrations of the enzymes lipoprotein lipase and hepatic lipase from the vascular endothelium than UFH. The clinical significance of this is unclear. Debate continues about whether LMWH should be regarded as a generic drug or whether each product should be treated as a separate entity (Prandoni & Nenci, 2003). As indicated in Table I, there are clearly recognisable differences in the in vitro properties of the various products. For regulatory purposes each manufacturer has to produce specific data on the pharmacology, toxicology and clinical effectiveness of a LMWH. However, there are a number of reasons for considering LMWHs as a family of closely related drugs. They share the same mechanism of action, and although produced by different chemical methods they have similar physicochemical properties. The differences in MW and anticoagulant activity seen in vitro are likely to be of diminished significance in vivo for two reasons. First, the molecules with high affinity to antithrombin tend to have a higher MW distribution than the low-affinity molecules. Secondly, after subcutaneous injection there is a filtering effect, whereby the highest MW molecules, which have the highest anti-IIa activity, are absorbed least. Thus, the MW and anticoagulant activities of the active species circulating after injection of the various products are likely to be much more similar than would appear from consideration of their in vitro properties. From a clinical point of view, the evidence published so far indicates that any differences in effectiveness and safety among the products, if they exist, must be extremely small, although there have been very few direct comparisons. However, this conclusion may only hold for the group of relatively similar LMWHs. Products, such as Bemiparin, which has a lower MW distribution and much lower anti-IIa activity than the others (Table I), and fondaparinux, the synthetic pentasaccharide with only anti-Xa activity and no anti-IIa activity and a longer half-life of 17 h, could conceivably demonstrate clinical differences; further studies are needed if such differences relative to the LMWHs with higher mean MW are to be substantiated. In the United Kingdom LMWHs have replaced UFH as the preferred option in many clinical situations. For example, LMWH therapy is now considered the treatment of choice for the prevention and treatment of venous thromboembolism (VTE) and treatment of acute coronary syndromes in most patients. A further recent development has been the introduction of fondaparinux for the prevention of VTE in patients with hip fracture and those having total knee or hip replacements (Heit, 2002). When deciding on which heparin preparation and what dose to use, clinicians must consider for each patient episode. The patient haemostatic potential and hence the intrinsic patient risk of thrombosis or bleeding (patient risk). The risk of thrombosis and bleeding associated with the procedure or condition of the patient (disorder risk). The relative efficacy of different heparin preparations and doses and the relative bleeding risk associated with these (heparin risk). The recommendations in this BCSH guideline generally reflect those of the ACCP and SIGN and are updated where appropriate to encompass recent studies. The clinician should refer to the latest edition of the British National Formulary for guidance on product information for licensed indications and dosing regimens of each heparin preparation in each situation. In some instances the antithrombotic superiority of LMWHs over UFH has not been proven but the lower risk of heparin-induced thrombocytopenia with thrombosis (HITT) generally favours their use over UFH (Warkentin et al, 1995). In view of the uncertainty of whether LMWHs are interchangeable, generic recommendations have been made in this guideline. Patients should be assessed for risk of VTE and given prophylaxis according to the degree of risk (Thromboembolic Risk Factors (THRIFT) Consensus Group, 1992). Both heparin and non-pharmacological methods should be considered as combined modalities are most effective (Geerts et al, 2004). Subcutaneous UFH at a dose of 5000 units 8–12 hourly reduces the risk of symptomatic deep vein thrombosis (DVT) and pulmonary embolism (PE) and death because of PE in patients having major general and gynaecological surgery (Collins et al, 1988; Clagett & Reisch, 1998). LMWHs are at least as effective and whilst bleeding rates are similar (Koch et al, 1997, 2001) there is a lower risk of heparin-induced thrombocytopenia and LMWHs can be administered by once daily subcutaneous injection (Warkentin et al, 1995). Patients undergoing major non-orthopaedic surgery should be considered for LMWH or UFH at recommended prophylactic dose (grade A). The value of thromboprophylaxis in orthopaedic surgery is clear (Collins et al, 1988). Collins et al's (1988) meta-analysis showed that UFH resulted in a statistically significant two-thirds reduction in all three outcomes of efficacy in the orthopaedic surgery subgroup: subclinical and clinical DVT, clinical PE and fatal PE. This meta-analysis also showed a 21% reduction in total mortality for all surgical patients primarily because of reduction in fatal PE. LMWHs have been compared with UFH in numerous studies and several meta-analyses have shown at least equivalence in safety and efficacy and in some a small but significant advantage in efficacy (Nurmohamed et al, 1992; Koch et al, 1997). Following total hip replacement the incidence of fatal PE is at least 0·37% as shown in the Norwegian Hip Arthroplasty Registry of 67 548 patients (Lie et al, 2002). This figure applied to patients who were receiving standard thromboprophylaxis with a LMWH. Unprotected patients would be expected to have a mortality rate of around three times that frequency. Prospective registries confirm that the incidence of fatal PE may have declined over recent years. However, this decline is not to the degree found in British studies, which have been retrospective and subject to incomplete follow up and data collection and are therefore contentious and not accepted internationally (Murray et al, 1996; Gillespie et al, 2000). Aspirin has also been assessed in the context of lower limb orthopaedic surgery. The PEP study reported that 160 mg of aspirin started preoperatively and continued for 35 d after hip fracture reduced the incidence of the secondary outcome of VTE by approximately one-third (Pulmonary Embolism Prevention (PEP) Trial, 2000). However, aspirin had no effect on the primary outcome of this study, major vascular events and vascular mortality. Aspirin also had no effect on VTE in hip or knee replacement surgery. The analysis and interpretation of the trial data have been criticised and strong recommendations have been made against aspirin use (Cohen & Quinlan, 2000; Geerts et al, 2004). Despite this, many orthopaedic surgeons now use suboptimal therapy with aspirin rather than a LMWH or UFH. There has been no direct comparison of aspirin with either UFH or a LMWH. Scottish Intercollegiate Guidelines Network concluded in 2002 that patients having total hip or knee replacement (or other elective major orthopaedic surgery) could be considered for UFH, LMWH or aspirin. The aspirin recommendation has recently been strongly discouraged (Geerts et al, 2004). Although this was presented as a grade A recommendation, the advice was that 'treatment could be considered'. Furthermore, no preference was indicated between UFH, LMWH and aspirin (http://www.sign.ac.uk/guidelines/fulltext/36/section12.html). In contrast, the 7th Consensus Conference of the American College of Chest Physicians on Antithrombotic Agents concluded that subcutaneous LMWH is the preferred prophylactic option for both elective hip and knee replacement (Hirsh & Raschke, 2004). For hip replacement, adjusted dose UFH was considered an acceptable but more complex alternative and aspirin was considered to be less effective and was not recommended. UFH and aspirin were not recommended for knee replacement. The optimal duration of prophylaxis after hip or knee replacement remains uncertain. Nowadays hospitalisation is usually for <5 d but DVT risk may persist for up to 2 months after hip replacement. Extended prophylaxis (usually 5 weeks) reduces the incidence of asymptomatic total and proximal DVT and symptomatic VTE by at least 50% (Cohen et al, 2001). The ACCP recommended LMWH prophylaxis for at least 7–10 d after surgery with extended prophylaxis for high-risk patients. In randomised-clinical trials, postoperative administration of fondaparinux has been shown to reduce the risk of asymptomatic VTE in patients undergoing major elective orthopaedic surgery compared with preoperative LMWH (Turpie et al, 2002). The rate of 'clinically relevant bleeding' (defined as leading to death or re-operation, or into a critical organ) was said to be not increased. However, there was a significant excess of 'major bleeding' in the fondaparinux group and an indication that excessive bleeding occurred in patients receiving fondaparinux <6 h after surgery. Subgroup analyses later allowed clinical and regulatory approval for commencing fondaparinux at least 6 h postoperatively. Fondaparinux may therefore be superior to LMWH for the prevention of VTE in this group of patients when given as recommended, but some uncertainties have still to be resolved (Heit, 2002). Fondaparinux does not appear to cross-react with anti-PF4/heparin antibodies. Patients undergoing major elective orthopaedic surgery should be considered for LMWH (or fondaparinux) at recommended prophylactic dose for at least 7–10 d (grade A). Venous thromboembolism rates and risk of fatal PE are higher in hip fracture patients than those having elective total hip or knee replacements. Heparins (UFH and LMWH) reduce the risk of asymptomatic VTE but there are insufficient data to establish the effect on symptomatic VTE or mortality (Handoll et al, 2002). Aspirin appeared to reduce the risk of VTE in the PEP study; however, there was no effect on mortality. Fondaparinux was shown to be more effective than a LMWH in a single randomised study (Eriksson et al, 2001). A more recent study has shown the benefit of prolonged therapy with fondaparinux resulting in a highly significant reduction in both asymptomatic and clinical outcomes (Eriksson & Lassen, 2003). Scottish Intercollegiate Guidelines Network has recommended that all patients with hip fracture should receive aspirin unless contraindicated (http://www.sign.ac.uk/guidelines/fulltext/36/section12.html). The 7th Consensus Conference of the American College of Chest Physicians on Antithrombotic Agents recommended LMWH for patients having hip fracture surgery. Aspirin was not recommended as it was considered less effective. Thromboprophylaxis with LMWH (or fondaparinux) at recommended prophylactic dose should be considered for hip fracture patients (grade A). At present there is no evidence that heparins reduce the risk of symptomatic VTE or fatal PE in trauma patients. A meta-analysis of UFH did not indicate benefit (Upchurch et al, 1995). A LMWH reduced the risk of asymptomatic VTE in trauma patients compared with UFH (Geerts et al, 1996). Thromboprophylaxis with LMWH at recommended prophylactic dose should be considered for major trauma when not contraindicated by bleeding risk (grade B). Low-molecular weight heparins have been shown to reduce the incidence of asymptomatic DVT in outpatients with lower limb plaster casts (Koch et al, 1995; Jorgensen et al, 2002; Lassen et al, 2002). An effect on fatal PE has not been established. Patients considered to be at high risk of VTE associated with lower limb plaster cast immobilisation may be considered for thromboprophylaxis with LMWH at recommended prophylactic dose (grade B) Because of the particularly serious consequences of surgery-associated bleeding in neurosurgical patients the antithrombotic role of heparin has been less well defined. Whilst UFH and LMWHs do reduce the risk of VTE there is a significant risk of major bleeding (Iorio & Agnelli, 2000) and for this reason mechanical methods of thromboprophylaxis may be preferable. When assessing the most suitable method of thromboprophylaxis consideration should be given to the relative risks from bleeding in relation to whether cranial surgery is intracranial or extracranial and whether spinal surgery is intradural or extradural. The principles of risk assessment apply to other types of invasive procedure and the clinician is referred to the Scottish Intercollegiate Guideline 62: prophylaxis of VTE (http://www.sign.ac.uk/guidelines/fulltext/62/index.html) for further recommendations in specific subgroups of patients, including those having spinal and epidural anaesthesia. Most cases of VTE occur in non-surgical patients. However, there are fewer randomised interventional studies with heparin and individual patient risk assessment is less structured and validated than in surgical patients. Meta-analysis has shown that heparin significantly reduces the risk of symptomatic VTE but there is no proven antithrombotic advantage of LMWHs over UFH (Mismetti et al, 2000). However, a lower bleeding risk and a lower risk of HITT favours the use of LMWH (Mismetti et al, 2000). Validated risk-assessment tools are required to determine which categories of medical patients should be routinely offered LMWH thromboprophylaxis. For example, in the Medenox study, patients more than 40 years old with an anticipated hospital stay of at least 6 d and either congestive heart failure or acute respiratory failure, or a medical condition with an additional risk factor for VTE (as specified in the inclusion criteria), were randomised to enoxaparin or placebo. Treatment with 40 mg enoxaparin was associated with a 63% reduction of venographically documented DVT or PE. Enoxaparin 20 mg was ineffective. More recently, the Prevention of Recurrent Venous Thromboembolism (PREVENT) study, a randomised, prospective, double-blinded study analysed the efficacy and safety of 5000 IU dalteparin compared with placebo as thromboprophylaxis in a total of around 3700 moderate risk hospitalised medical patients for a total of 14 d. Patients were then assessed for the presence of asymptomatic proximal DVT and a significant 45% (P = 0·0015) reduction in VTE was observed in the treatment group (Leizorovicz et al, 2004). In addition to this, the ARTEMIS study has evaluated the use of the factor Xa inhibitor, fondaparinux, in thromboprophylaxis of medical patients. A significant 47% (P = 0·029) reduction in VTE and in fatal PE was seen (Cohen et al, 2003). Medical patients determined to be at high risk of VTE should be considered for thromboprophylaxis with LMWH at recommended prophylactic dose (grade A). Retrospective meta-analysis of heparin trials suggests that patients with cancer treated with LMWHs for VTE have a survival advantage. This observation has now also been made in some small prospective studies in which patients without VTE were randomised to a LMWH or placebo, in addition to chemoradiotherapy (Altinbas et al, 2004; Kakkar et al, 2004). The mechanism remains unclear and long-term benefit has not yet been proven. Future studies are required to confirm a beneficial effect and address issues, such as patient selection, dose and duration of therapy. At this stage we do not recommend that patients should receive heparin as an antineoplastic agent outside clinical trials. However, many hospitalised patients with cancer will fall into a high-risk group for VTE and should be considered for thromboprophylaxis with LMWH. The randomised comparison of LMWH versus oral anticoagulant therapy for the prevention of recurrent VTE in patients with cancer (CLOT) study (Lee et al, 2003) indicated that secondary thromboprophylaxis with a LMWH was more effective than with oral anticoagulant in a cohort of patients with VTE and cancer. Heparins are not recommended for use as antineoplastic agents outside clinical trials. Antithrombotic therapy has not yet been shown to alter the incidence or severity of sickle cell crisis. The incidence of VTE during sickle cell crisis is unclear. The BCSH General Haematology Task force has given an ungraded recommendation that prophylactic anticoagulation should be considered for all patients confined to bed for more than 16 h/d, particularly if there are additional risk factors for VTE, such as a history of previous thromboembolism or insertion of a femoral line. A LMWH at prophylactic dose would seem reasonable. Sickle cell disease is considered an additional risk factor for pregnancy-associated VTE and should be taken into consideration in assessing the need for thromboprophylaxis in pregnancy (http://www.rcog.org.uk/index.asp?PageID=533). Patients in sickle cell crisis should be considered for LMWH at recommended prophylactic dose until recovery from crisis (grade C). Heparins, including LMWHs do not cross the placenta and are the anticoagulant of choice for prevention and treatment of pregnancy-associated VTE. Randomised trials of prophylaxis and treatment have not been performed specifically in pregnant women. Guidelines for thromboprophylaxis have been published by the BCSH (Walker et al, 2001) and for treatment by the Royal College of Obstetricians and Gynaecologists (RCOG; http://www.rcog.org.uk/index.asp?PageID=533). The principal considerations are: due to the altered pharmacokinetics a twice daily dosing regimen for LMWHs is recommended; frequent anti-Xa monitoring is not recommended but, if possible, anti-Xa activity should be measured to confirm appropriate dosing, at least until more information is available on the therapeutic use of LMWH in pregnancy (but see below regarding the limitations of monitoring); the platelet count should be monitored in the early stages of administration to avoid delayed diagnosis of heparin-induced thrombocytopenia. However, this complication is exceedingly uncommon when LMWH is used prophylactically in asymptomatic pregnant women and in this situation monitoring is not requried; the duration of therapeutic anticoagulation in the non-pregnant subject with VTE is usually 6 months. As pregnancy is associated with prothrombotic changes in the coagulation system and venous flow, and as the increased coagulation activation persists for some weeks after delivery, a similar duration of anticoagulation is prudent in pregnancy VTE. Thus, therapeutic anticoagulation should usually be continued for at least 6 months. If the VTE occurs early in the pregnancy, provided that there are no additional risk factors, reduction of the dose of LMWH or UFH to prophylactic levels could be considered after 6 months of treatment; and the woman should be advised that once she is established in labour or thinks that she is in labour, she should not inject any further heparin. She should be reassessed on admission to hospital and further doses should be prescribed by medical staff. Decisions regarding the use of spinal anaesthesia in relation to heparin administration should be guided by the perceived benefits in the individual case balanced against the potentially catastrophic effects of local bleeding and should be made by an experienced anaesthetist. Heparin-associated skin reactions appear to be more common in pregnant women. This may relate to the unusual duration of heparin exposure. Cross-reactivity between heparin preparations, including LMWHs is common and several different heparins may have to be tried. Ovarian hyperstimulation syndrome during assisted reproduction may be complicated by arterial or VTE. There are insufficient data to reach conclusions on the efficacy of heparin thromboprophylaxis in this situation. In vitro data suggest fondaparinux is also unlikely to cross the placental barrier (Lagrange et al, 2002), but there is very little experience of its use in pregnancy to date. Heparin has been shown to reduce the risk of fatal recurrence in patients with symptomatic PE (Barritt & Jordan, 1960) and to result in a low risk of recurrent non-fatal VTE (Carson et al, 1992; Douketis et al, 1998). LMWHs are at least as effective as UFH for treatment of submassive PE and DVT (Gould et al, 1999). LMWHs are preferable in view of the lower risk of HITT. Patients with massive PE should be considered for treatment with thrombolytic therapy (British Thoracic Society Standards of Care Com