prevention, diagnosis and treatment
Rhona M. Maclean
Venous thromboembolism (VTE) is the most common cause of preventable death in hospitalised patients, with an estimated 25 000 patients dying from preventable hospital acquired VTE in the UK each year.1 The two most common manifestations of venous thrombosis are deep vein thrombosis (DVT) and pulmonary embolism (PE), both of which are associated with significant morbidity (post-thrombotic syndrome and chronic thromboembolic pulmonary hypertension respectively). As surgery is associated with a high risk of postoperative VTE, evidence-based thromboprophylaxis strategies should be employed to reduce this risk, and those patients with symptoms and/or signs suggestive of VTE should be thoroughly investigated and managed accordingly.
Epidemiology of VTE
The incidence of a first episode of VTE is estimated at 1–2 per 1000 person-years in white Caucasians, with a lower incidence in Hispanics and Asians.2,3 The risk of VTE increases progressively with age, with an incidence of > 5/1000 person-years in those over the age of 80.3 The incidence of VTE is approximately equal in men and women; however, it is more frequent in women in the childbearing years, likely due to the use of the hormonal therapies (the combined oral contraceptive pill and hormone replacement therapy) and pregnancy, whereas after the age of 45 incidence rates are generally higher in men. Clinical studies, excluding autopsy data, have consistently shown that the incidence of DVT is approximately twice that of PE.2,3 It has been estimated that there are over 130 000 cases of VTE in the UK each year, at a cost (direct and indirect) of around £640 million annually.4
DVT and PE are the result of the same disease process, VTE; of patients presenting with symptomatic PE, up to 80% will have asymptomatic DVT, and in those presenting with symptomatic DVT, 50–80% can be shown by imaging to have PE.5
Approximately half of VTE episodes are idiopathic in nature (defined as having had no recent surgery, trauma, cancer, pregnancy or immobilisation), the remainder presenting after an obvious precipitating event.2,3 Of patients diagnosed with VTE, the majority (73.7%) present as outpatients and of these a significant proportion have either undergone surgery (23.1%) or been hospitalised (36.8%) in the preceding 3 months.6 Presentation with VTE after surgery peaks at 21 days postoperatively, with the risk remaining significantly increased for 3 months after surgery.7
VTE is associated with a surprisingly high mortality. In a retrospective epidemiological study, the 30-day fatality rate was 4.6% following DVT and 9.7% with PE.3 Cancer was associated with a worse outcome, with a 30-day fatality rate of 19.1%. Mortality rates associated with PE are high, with approximately 25% of patients dying within a year, many due to malignancy, others from cardiopulmonary disease or recurrent VTE.7 Overall, pulmonary emboli are thought to be responsible for 10% of hospital inpatient deaths,8 with many diagnosed for the first time at post-mortem.9
Thrombosis in the deep veins damages the deep venous valves, resulting in post-thrombotic scarring, stiffening of the vessel wall, venous insufficiency, reflux and venous hypertension, all of which can result in swelling, pain and heaviness in the affected leg – the post-thrombotic syndrome. In its severe form, this causes marked skin changes – lipodermatosclerosis (varicose eczema and atrophy of the subcutaneous tissues), hyperpigmentation and ulceration – and can be a major cause of morbidity. The incidence of the post-thrombotic syndrome has been reported at approximately 28% after 5 years, with 9% having a severe form.7 Up to 5% of patients with PE will develop pulmonary hypertension.7
Pathophysiology of venous thromboembolism
DVTs usually start in the calf veins, and thrombi developing after surgery often originate in the valve cusps of the soleal veins. These thrombi are initially formed of red blood cells caught in a fibrin mesh, which then incorporate platelets and fibrin into the clot, propagating proximally to form a free-floating thrombus, or extending to occlude the vein.10,11 Many such thrombi start to develop intraoperatively; of these, half will resolve spontaneously within 72 hours and 18% will extend proximally, of which 50% will embolise. Some thromboses, however, begin postoperatively; 20–34% of patients diagnosed as having DVT by screening tests in hospital had been shown to have legs free of thrombosis postoperatively. It has been estimated that 10% of symptomatic PEs cause death within 1 hour of onset (Box 14.1). There is an ongoing debate as to the significance of calf vein thrombosis and whether, if detected, it should be treated. A recent study of isolated symptomatic muscular calf vein thromboses suggested that they are not as innocuous as originally thought; 7% of patients had symptomatic PEs at presentation, and after completing 1–3 months of anticoagulation, 18% had recurrent episodes of VTE within 3 years.12
Box 14.1 Natural history of venous thromboembolism
In the 1860 s, Virchow described three factors implicated in the development of a venous thrombosis: slowing of venous blood flow (stasis), damage to the vessel wall (venous injury), and changes in the blood that increase the propensity to develop thrombosis (hypercoagulability). These are now widely known as Virchow's triad, and this model remains a valid concept today.
Stasis, the slowing of venous return, is associated with an increased risk of VTE (stroke patients have an increased risk of DVT in the paralysed limb13). It has been demonstrated that intraoperative paralysis causes ‘microtears’ in the venous endothelium, exposing circulating blood to procoagulant components in the subendothelium (e.g. collagen, von Willebrand factor, tissue factor).14 Stasis, in and of itself, is likely to be insufficient to cause VTE, and most patients with significant immobility are likely to have a systemic illness, increasing the risk of VTE by other mechanisms.
Surgery or trauma cause venous injury, which increases the risk of venous thrombosis. Inflammatory cytokines also induce venous injury by down-regulating thrombomodulin, impairing fibrinolysis, stimulating tissue factor expression on monocytes or endothelial cells, and inducing apoptosis of endothelial cells, rendering them thrombogenic.15
Hypercoagulability, both inherited and acquired, has been demonstrated to significantly influence the development of VTE. Studies undertaken in families in whom multiple members have presented with VTE have identified a number of inherited thrombophilias.
Risk factors for VTE
Antithrombin (AT) is an anticoagulant protein, whose function is predominantly to inactivate thrombin and activated coagulation factor X, thereby limiting ongoing thrombus propagation. AT deficiency is found in 1% of consecutive patients presenting with VTE and increases the risk of VTE five- to 50-fold.16,17 Acquired AT deficiency can occur in a number of different clinical situations, including liver disease, sepsis, acute thrombosis, disseminated intravascular coagulation (DIC) and nephrotic syndrome.
Proteins C and S are both vitamin K-dependent proteins; their activity will be reduced by vitamin K antagonist anticoagulants(such as warfarin and sinthrome). The activated protein C/protein S complex inactivates coagulation factors V and VIII, also limiting ongoing thrombus propagation. Inherited protein C deficiency increases the risk of VTE six- to 15-fold and is found in 1–3% of patients presenting with VTE.16 Protein S deficiency is present in 1–3% of patients with VTE. Liver disease, sepsis, DIC and acute thrombosis reduce the levels of proteins C and S, resulting in acquired hypercoagulability.
Factor V Leiden and the prothrombin gene mutation
Factor V Leiden, a mutation of the factor V gene, is the most common inherited thrombophilia, present in 5% of the Caucasian population (and < 1% of Africans/South East Asians). It renders activated factor V relatively resistant to inactivation by the activated protein C/protein S complex, and confers an eightfold increased risk of VTE in the heterozygous form (80-fold increased risk in the homozygous form). In Caucasian populations, factor V Leiden is found in > 20% of unselected patients presenting with VTE.16 The prothrombin gene mutation (G20210A) is less prevalent, found in 1% of the general population, and confers a threefold increased risk of VTE. It is found in 5–6% of unselected patients with VTE.16
Classical homocystinuria causes extremely high levels of plasma homocysteine and is associated with both venous and arterial thrombosis in addition to the other classical disease manifestations (mental retardation, seizures, musculoskeletal abnormalities, eye anomalies including lens dislocation). Neither the common mutation in the methylene tetrahydrofolate reductase gene (MTHFR C677T) nor mild elevations in homocysteine levels are associated with thrombosis.18
Testing for inherited thrombophilia, whilst frequently undertaken, rarely helps in the management of patients with VTE. Such testing should usually only be performed in young individuals with a personal history of VTE who have a history of VTE in a first-degree relative. Testing should not usually be done while a patient is taking anticoagulant therapy.19
Elevated factor VIII levels have been demonstrated to increase the risk of developing a first VTE, and also increase the risk of recurrent thrombosis.16,20
Association and linkage studies have identified a number of other proteins associated with VTE. This includes high levels of the coagulation factors fibrinogen, prothrombin, IX, XI and von Willebrand factor. There is also an association between VTE and platelet glycoprotein VI, blood group O and deficiencies of proteins associated with fibrinolysis (plasminogen and plasminogen-activated inhibitor-1 (PAI-1)). These proteins are associated with a weak increased risk of thrombosis (1.1- to 2.5-fold).16,21,22
Antiphospholipid syndrome (APLS)
Diagnosis of the APLS requires the presence of both clinical (arterial or venous thrombosis or recurrent miscarriage) and laboratory (persistent detectable antiphospholipid antibodies such as lupus anticoagulant or anticardiolipin or anti-β2-glycoprotein 1 antibodies) criteria. The mechanism of these antibodies in the development of thrombosis has not been established; however, a number of hypotheses have been suggested, including the interference of antibodies with anticoagulant mechanisms, triggering procoagulant changes in leucocytes, platelets or endothelial cells, or activation of complement triggering an inflammatory reaction.23 Patients with a diagnosis of APLS who have had a thrombosis should be therapeutically anticoagulated long term (usually with warfarin, target international normalised ratio (INR) 2.5).23
Heparin-induced thrombocytopenia (HIT)
HIT is a rare but life-threatening complication of heparin therapy; it confers a high risk of thrombosis (30–75%), and has a significant morbidity and mortality. It occurs more frequently in patients receiving unfractionated heparin (UFH) than low-molecular-weight heparin (LMWH), and the highest risk is in those who have had cardiothoracic surgery. General surgical patients receiving LMWH have a < 1% risk of developing HIT. The platelet count characteristically falls by ≥ 50% from baseline (rarely below 20 × 109/L) between days 5 and 10 of heparin therapy. Less often, if a patient has received heparin within the last 100 days, HIT can present acutely after heparin administration with systemic symptoms (rigors, cardiorespiratory distress). Skin lesions at the site of heparin injections have also been shown to be associated with HIT.24 Half of patients with HIT will develop thrombosis, and unless heparin is stopped and alternative anticoagulation commenced, there is a considerable risk of further thrombosis developing.25 HIT is rarely associated with bleeding and so platelet transfusions should not be given due to the risk of thrombosis,25 unless there is active bleeding.
Patients receiving heparin should have a platelet count checked on the day treatment is started, after 24 hours of therapy (if exposed to heparin within the previous 100 days) and thereafter every 2–4 days until day 14. If the platelet count falls by ≥ 50% of baseline or the patient develops new venous or arterial thrombosis or skin allergy, then a diagnosis of HIT should be considered and a clinical assessment undertaken; advice should be sought from haematology. If the diagnosis of HIT is thought to be likely, heparin should be stopped (including heparin flushes) and, due to the high risk of thrombosis, an alternative anticoagulant (e.g. danaparoid, argatroban or fondaparinux) should be commenced, pending the results of further investigations. Again, in such circumstances the advice of a haematologist is essential. Warfarin should not be started until the platelet count has fully recovered, and care should be taken to continue an additional anticoagulant until the INR is > 2.0 for 2 days.
Other Risk Factors For VTE
Age is an important risk factor for VTE, with many studies showing an increased risk in patients over 40 years of age and considerably greater in those > 70 years of age.26 It is likely that this is a reflection of medical comorbidities, immobility and coagulation activation.
Those with a body mass index (BMI) of over 30 kg/m2 have a two- to threefold increased risk of VTE. As with age, it is thought that this might be a reflection of immobility and coagulation activation.26
Family history of VTE
A history of VTE in a first-degree family member (aged < 50 years) confers an increased risk of developing a VTE.27
Medical illness and malignancy
Hospitalisation itself is associated with an eightfold increased risk of VTE, with specific patient groups being at particularly high risk. Congestive cardiac failure, acute infection, central venous access, paralytic stroke, nephrotic syndrome, cancer and chemotherapy are all moderate risk factors for VTE (odds ratio 2–9).28 It has therefore been proposed that hospitalised medical patients are risk assessed for their level of risk of developing VTE (NICE guideline,29 SIGN guideline,30 ACCP guideline8).
Cancer is strongly associated with VTE; 15% of patients with VTE have malignancy and 2–3% of patients with VTE are newly diagnosed with malignancy at presentation. Patients with certain cancers (stomach, lung, breast, pancreas, gynaecological, lymphoma) are at particularly high risk of thrombosis, thought to be caused by tissue factor-like substances and microparticles activating the coagulation cascade. Surgery and immobilisation further increase this patient group's risk of thrombosis.
Hormone replacement therapy (HRT) and combined oral contraceptive pill (cOCP)
The risk of VTE is increased two- to fourfold in users of the cOCP.31 The risk is highest in the first year of use, diminishes thereafter, and is reduced by the use of the lower dose oestrogen preparations. Obese women (BMI > 30) have a twofold increased risk of VTE, but a 10-fold increased risk if taking the cOCP. The risk of VTE after surgery in women taking the cOCP is increased 2.5-fold. There is no evidence that the progesterone-only contraceptive increases the risk of VTE. Women with thrombophilia (antithrombin, protein C or S deficiencies, factor V Leiden of the prothrombin gene variant) who take the cOCP are at particularly high risk of VTE.32 More recently, transdermal contraceptive patches have been introduced containing both oestrogen and progestogen, but the risk of VTE with these preparations appears to be similar to the oral preparations.33
HRT is associated with a two-to fourfold increased risk of VTE in women using HRT compared with non-users.32–34 As with the cOCP, the risk is highest in the first year of HRT,35 and there is a synergistic effect with the inherited thrombophilias.36 Unlike the cOCP, there appears to be a lower risk of VTE when the transdermal route is used compared with oral preparations.37
Pregnancy and puerperium
VTE is a leading direct cause of maternal death in the UK, primarily due to pulmonary embolism. There is a 10-fold increased risk of VTE in pregnancy and a 25-fold increased risk during the puerperium. Many of these thrombotic events are preventable by the use of appropriate thromboprophylaxis38 and it therefore follows that all pregnant women, including those in the postpartum period, should be considered ‘at risk’ of VTE if admitted with an acute illness, and given thromboprophylaxis unless contraindicated. Low-molecular-weight heparins are safe to use in pregnancy as they do not cross the placenta into the foetal circulation.
Travel of long duration is a relatively weak risk factor for VTE, with risks higher in those with pre-existing risk factors. Studies that have investigated the role of risk factors in travel-related thrombosis have mentioned the role of recent trauma or surgery, in addition to obesity, cancer and hormone therapy. It is important to remember that the risk of developing symptomatic VTE after long-duration travel remains increased in the 8 weeks after the flight.39–41
Superficial thrombophlebitis (STP) of the lower leg is a significant risk factor for VTE and a large prospective epidemiological study found that 3.3% of patients with STP develop symptomatic VTE if untreated. Those with STP > 5 cm in length were more likely to have associated DVT if the STP was in the proximal long saphenous vein. STP within a varicose vein was less likely to be associated with DVT.42
One-third of patients with VTE will have had surgery in the preceding 3 months, with the risk greatest following major abdominal and pelvic surgery (especially if associated with malignancy), and major orthopaedic procedures. Without appropriate thrombosis prevention strategies, 60–80% of such patients will develop venous thrombosis.8 Major trauma is also associated with a very high risk of VTE. It has been recognised that, in addition to the surgical procedure itself, a number of other factors increase the risk of thrombosis in the surgical patient (Box 14.2). It is also now understood that the risk of VTE does not end at the point of discharge as 56% of all VTE episodes within 91 days of surgery occur after discharge.43,44
Box 14.2 Risk factors for venous thromboembolism
Protein C deficiency
Protein S deficiency
Factor V Leiden
Elevated factor VIII levels
Acute medical illness
Cancer therapies (hormonal, chemotherapy or radiotherapy)
Pregnancy and puerperium
Oestrogen-containing oral contraceptives or hormone replacement therapy
Selective oestrogen receptor modulators
Heart or respiratory failure
Inflammatory bowel disease
Paroxysmal nocturnal haemoglobinuria
Varicose veins and superficial vein thrombophlebitis
Central venous lines
Prevention of venous thromboembolic disease
Over the last few years there has been a proliferation of guidelines for the prevention of VTE, with national guidelines developed in Scotland (SIGN 122)30 and England and Wales (NICE CG92).29 All hospitals should formally adopt and implement guidelines for VTE prevention and audit their practice.
Studies investigating strategies for the prevention of VTE have used angiography or more recently ultrasonography to confirm the presence of a thrombotic event, with the majority of these being subclinical below-knee DVTs. There is a strong association between asymptomatic DVT and the development of symptomatic VTE (proximal DVT or PE)8 and it is therefore essential to introduce thromboprophylaxis strategies in patients at risk in order to reduce the morbidity and mortality associated with this disease. While it is possible to identify groups at high risk of VTE,45 it is not possible to identify the specific patient who will develop a symptomatic thrombosis, nor the individual who will develop a sudden massive PE. In 70–80% of patients dying of a PE in hospital, a diagnosis of VTE was not considered prior to autopsy, and most patients with PE have had no previous symptoms suggestive of DVT.
The prevalence of postoperative PE is now considerably lower than it was some years ago due to the use of improved surgical and anaesthetic techniques, early mobilisation of patients and the systematic introduction of chemical thromboprophylaxis.
Methods of thromboprophylaxis
Mobilisation And Leg Exercises: While evidence to support early mobilisation and leg exercises in reducing the risk of VTE is scarce, immobility significantly increases the risk of VTE, and trials of bed rest for medical illnesses demonstrated no evidence of benefit.26 Mobilisation and leg exercises are simple interventions and should be encouraged where appropriate to reduce thrombosis risk.
Mechanical Thromboprophylaxis: Mechanical methods of prophylaxis include graduated elastic compression stockings (GECSs), intermittent pneumatic compression (IPC) devices and venous foot-pumps. These methods are designed to reduce venous stasis and improve venous flow in the leg veins.
GECSs have a compression profile of 18 mmHg at the ankle, 14–15 mmHg mid-calf and 8 mmHg at the knee. They are simple, safe when applied correctly (care should be taken not to use them in patients with severe peripheral vascular disease) and moderately efficacious in preventing VTE without any increased risk of bleeding.46
Graduated compression stockings produce a highly significant risk reduction in postoperative venous thomboembolic disease of 68% in moderate-risk patients.47 A recent Cochrane review confirmed the effectiveness of GECSs in reducing the risk of venous thromboembolic disease, particularly when used on a background of other methods of prophylaxis.48 All surgical patients at risk of VTE should have GECSs applied unless contraindicated.
There are a number of different manufacturers of IPC devices, which supply intermittent low pressure (typically 40 mmHg) for 1 minute with a 90-second period of decompression. This increases the velocity of venous return and blood flow and may also enhance fibrinolysis49 and inhibition of the tissue factor pathway.50 A meta-analysis47 of the use of mechanical methods for thromboprophylaxis reported DVT rates of 9.9%and 17.6% for IPC versus 20.3% and 27% for placebo. Adherence with IPC is frequently less than optimal in ‘real-life’ settings.
A major limitation in the studies of intermittent pneumatic compression (IPC) devices is the relatively small numbers of patients included. When used alone in patients undergoing high-risk surgery they appear to lack efficacy, and further assessment of their role is required.
Vitamin K Antagonists: It has been known since 1959 that warfarin anticoagulation significantly reduces the incidence of thromboembolic disease when used for prophylaxis.51 However, this method is not routinely used, mainly because of the considerable risk of haemorrhage, either spontaneous or related to the surgery.
Aspirin And Antiplatelet Agents: The Medical Research Council sponsored a double-blind, placebo-controlled study of 303 surgical patients, which showed no reduction in postoperative DVT with 600 mg aspirin taken immediately before and for 5 days after surgery.52 A meta-analysis of thromboprophylaxis following total hip replacement, including data from 56 randomised trials, showed no benefit for aspirin in preventing DVT.53 However, in the meta-analysis of the Antiplatelets Trialists' Collaboration,54 aspirin was found to significantly reduce DVT, although the poor quality of some of the studies included has led to many doubts regarding this finding. Aspirin may have a role in prophylaxis of low-risk groups.54
From the results of these studies,52–55 aspirin appears to have some efficacy in the prevention of VTE, but as there are no robust studies directly comparing it with other pharmacological agents of proven efficacy, aspirin should not currently be considered a thromboprophylactic drug of choice.
Heparins: The use of low-dose unfractionated heparin (UFH) for thromboprophylaxis has now largely been superseded by low-molecular-weight heparin (LWMH) due to the latter's superior bioavailability, once-daily dosage and the reduced incidence of HIT. Heparins are renally excreted and therefore UFH may be preferred in patients in whom there is an increased risk of accumulation (those with significant renal impairment).
A meta-analysis of UFH prophylaxis in surgery showed a reduction in DVT and PE by two-thirds when compared to placebo, with a 2% increase in minor bleeding events.56 Studies have consistently shown reductions in DVT and PE by two-thirds with both UFH and LMWH, with a slightly reduced risk of bleeding with LMWH compared with UFH. Two meta-analyses have shown superior results with LMWH after total hip replacement with regard to postoperative DVT (risk reduction of 17–32%) and PE (risk reduction of 50%).57,58 Higher doses of LMWH should be given to high-risk general surgical patients as these provide greater protection than lower doses. In cancer patients undergoing surgery, prophylaxis with dalteparin 5000 units daily was shown to be significantly more effective than 2500 units daily, without an increased risk of bleeding.59 Compared with no prophylaxis, LMWH reduces the risk of clinical PE and clinical VTE by approximately 70%, and is associated with a possible reduction in the risk of death from any cause. However, LMWH also leads to an approximately doubling of the risk of wound haematoma and is associated with a small increase in major bleeding, which is higher in those with renal failure.
Fondaparinux: Fondaparinux is a synthetic pentasaccharide and a potent, highly selective, inhibitor of factor Xa. It has a longer half-life, is associated with a slightly increased risk of bleeding and is more expensivethan LMWH. For these reasons, fondaparinux is less commonly used for VTE prevention in general surgical patients. As LMWH is of porcine origin, fondaparinux can be used as an alternative in patients with religious objection to the use of porcine products.60
New Oral Anticoagulants: Dabigatran etexilate, an oral direct thrombin inhibitor, and rivaroxaban and apixaban (oral anti-Xa anticoagulants) have been licensed for thromboprophylaxis following hip and knee replacement surgery, but as yet there are few data for their use in general surgery.
Duration of thromboprophylaxis
Patients remain at increased risk of VTE once discharged after a surgical procedure, with a peak at 3 weeks postoperatively. Clinical trials of extended prophylaxis (4 weeks of LMWH vs. 1 week LMWH) after general surgery showed a significant reduction in venographically detected thromboses in those with cancer.8,61 One randomised trial of 300 patients undergoing abdominal or pelvic surgery receiving either 9 or 30 days of LMWH showed two proximal DVTs in those receiving short-term treatment, compared to one in those in the extended group.62
There are now many guidelines that provide recommendations on how to reduce the risk of VTE in surgical patients (and in other patient groups). In the UK these include the National Institute of Health and Clinical Excellence (NICE) Clinical Guideline 92 (Venous thromboembolism: reducing the risk29) and the Scottish SIGN 122 (Prevention and management of venous thromboembolism30). What these guidelines have in common is the recommendation that all patients should be assessed for their risk of developing a venous thrombosis based on both their individual predisposing factors and the risk associated with their illness and/or proposed surgical procedure. Patients assessed as being at high risk of VTE should be given thromboprophylaxis, which should be continued after discharge in specific high-risk groups.
Diagnosis of venous thromboembolic disease
Many patients present with leg pain or swelling, chest pain or breathlessness, and it is not possible to exclude VTE on the basis of clinical history and examination alone. Approximately 25% of patients with symptoms thought to be due to a DVT or PE will have those diagnoses subsequently confirmed; therefore, many could be unnecessarily exposed to imaging investigations with their associated risks. As a result, over the last 5–10 years, diagnostic strategies have been developed that utilise both clinical pretest probability scoring tools (for DVT and PE) and D-dimer estimation to reduce the requirement for imaging investigations.
Diagnosis Of DVT
Clinical features of DVT
The symptoms and signs of DVT tend to be non-specific. Most often patients complain of pain, swelling and redness of the lower limb, and on examination mild erythema, tenderness along the thigh or calf, muscle induration and mild pyrexia are found. However, many patients with DVT will have none of these clinical features. Unilateral pitting oedema is the most significant clinical sign, indicating thrombosis in 70% of patients. Homan's sign (pain or discomfort on foot dorsiflexion) is unreliable.
Patients with distal iliac or femoral vein thromboses can present with a very painful, swollen, white leg (phlegmasia caerulea alba). A rare but dramatic presentation is that of ‘phlegmasia caerulea dolens’, which occurs when the whole iliac system is thrombosed. The entire leg is swollen, acutely painful, dusky blue in colour and patchy areas of venous gangrene can develop. The cyanosis and swelling distinguish this from arterial ischaemia (in which the limb is white and rarely is there any swelling).
Venous thrombosis is not the only cause of a painful and/or swollen leg. Rupture of a Baker's cyst presents with sudden-onset lower leg pain, swelling and cellulitis, with a painful and erythematous lower limb. Muscular tears and bleeding into a muscle in a patient already on anticoagulant therapy can also be difficult to differentiate from a DVT. It is also important to remember that (rarely) malignancies such as sarcoma can present with a unilateral swollen leg.63
Diagnostic algorithms for DVT
Clinical prediction tools incorporate clinical symptoms, signs and risk factors of thrombosis, and assess the likelihood of a patient having a DVT. The most frequently used clinical prediction tool for DVT is the Wells score,64,65 which has been adapted to categorise patients as ‘DVT likely’ or ‘DVT unlikely’64 (Table 14.1). The Wells score is not validated for use in patients with previous DVT or in hospitalised or pregnant patients; initial investigation in such patients should be by appropriate imaging.
Clinical model predicting the pretest probability of DVT*
Active cancer (patient receiving treatment for cancer within previous 6 months or palliative)
Paralysis, paresis or recent plaster immobilisation of the lower extremities
Recently bedridden for 3 days or more, or major surgery within 12 weeks
Localised tenderness along the distribution of the deep venous system
Entire leg swollen
Calf swelling at least 3 cm larger than that on asymptomatic leg (10 cm below tibial tuberosity)
Pitting oedema confined to symptomatic leg
Collateral superficial veins (non-varicose)
Previously documented DVT
Alternative diagnosis at least as likely as DVT
*A score of 2 or higher indicates that the probability of DVT is likely; a score of less than 2 indicates that the probability of DVT is unlikely.
Reproduced from Wells PS, Anderson DR, Rodger M et al. Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis. N Engl J Med 2003; 349(13): 1227–35. Massachusetts Medical Society. All rights reserved.
D-dimers are the product of fibrinolysis of cross-linked fibrin, and are usually (but not always) increased in patients with active VTE (i.e. active thrombus formation). There are a number of different D-dimer assays (qualitative and quantitative) available, many of which have been evaluated to determine clinically meaningful cut-off values for the exclusion of VTE.66 A systematic review of patient cohorts presenting with suspected VTE who were prospectively assessed using a clinical prediction tool and D-dimer testing found a rate of VTE of 0.45% (95% confidence interval 0.22–0.83) in those who were at low risk of VTE and had D-dimers below the cut-off value.67 Patients who are considered using clinical prediction tools to be ‘DVT unlikely’, without increased D-dimers, can be considered to have had a DVT excluded and do not require further investigation. It must be borne in mind, however, that D-dimers tend to be increased with increasing age, and in those with malignancy, recent surgery or acute infection. The clinical utility of requesting a D-dimer assay in these patient groups is therefore limited, and most patients within these groups are likely to need imaging to exclude a thrombosis. D-dimer levels are reduced in patients on anticoagulant drugs.
Imaging Techniques For DVT
Venous ultrasound is the investigation of choice for suspected DVT, with high sensitivity (94–99%) and specificity (89–96%) for the diagnosis of symptomatic lower-limb proximal DVT when compared to contrast venography. Sensitivity and specificity are considerably less for calf DVT.68 The outcome of patients with a negative initial scan appears similar to control populations; however, there is evidence that distal DVT may extend proximally and subsequently embolise.69,70 For that reason, many centres repeat the ultrasound in 1 week if the initial scan is negative.
Care should be taken in the interpretation of ultrasound scans when requested in individuals with a history of previous DVT. A scan report of echogenic or non-occlusive thrombus at the site of a previous clot could be due to chronic or ‘old’ DVT rather than a new event. Serial ultrasound (or venography) may be useful in such circumstances.
Impedance plethysmography (IP)
IP was extensively studied in the 1970s and 1980s and is currently still used to exclude DVT in a number of centres. A tourniquet is applied to the affected limb and changes in the volume of the limb measured after tourniquet removal. The sensitivity and specificity of IP for DVT (83% and 92%, respectively) are lower than with ultrasound, hence the preference for ultrasound for the diagnosis of DVT by many hospitals.68,71
Venography was for a long time the ‘gold standard’ for a diagnosis of DVT; however, it is painful, involves the use of contrast medium and radiation, and is rarely used today. Venography has now been superseded by duplex ultrasound imaging.
Computed tomography (CT)
CT has a 95% sensitivity for DVT (proximal and distal) and 97% specificity.71 While it may not be routinely used for the diagnosis of peripheral venous occlusion, it is useful in the investigation of suspected proximal venous thrombosis (inferior and superior vena cava).
Magnetic resonance imaging (MRI) and angiography (MRA)
MRI and MRA can detect both peripheral and central venous thromboses, and are being studied for their role in the investigation of venous thrombosis. MRA is non-invasive, uses venous blood flow as the source of image contrast, but is not yet widely available. As with the other imaging modalities, MRA is sensitive and specific for proximal thrombosis, but has a lower sensitivity for distal thrombi. MRI is particularly useful for the investigation of suspected iliac thrombosis in pregnant women.
Summary Of Diagnostic Methods In DVT
Ultrasonography by an experienced operator is now the investigation of choice for DVT. D-dimer testing has now been widely incorporated into diagnostic algorithms, allowing the exclusion of DVT in patients at low risk with D-dimers below the cut-off for that assay. Serial ultrasound scanning should be performed in those at high risk of DVT with an initial ‘negative’ ultrasound but D-dimers above the cut-off. More invasive investigations (e.g. CT, MRA, venography) should be reserved for those in whom there is thought to be iliac thrombosis not visualised by ultrasound, or if there is lack of concordance between clinical probability and the ultrasound scan result. Most centres will prefer to use CT or MRA rather than venography for the investigation of patients suspected of having an iliac thrombosis.
Diagnosis Of Pulmonary Embolism
Clinical presentation and pretest probability
The symptoms and signs most commonly associated with a confirmed diagnosis of PE are (in order of descending frequency): dyspnoea, tachypnoea, pleuritic pain, apprehension, tachycardia, cough, haemoptysis, leg pain and clinical DVT. Clinical history and examination are insufficient to exclude a diagnosis of PE.72
As with DVT, the diagnostic process has been improved considerably by the introduction of clinical decision rules (such as the Wells (Table 14.2) and Geneva scoring systems), which risk stratify patients with suspected PE, and are used in combination with D-dimer testing. Using the PE Wells score and a quantitative D-dimer test to rule out PE, in those at low risk of PE with low D-dimers, 0–1.9 cases of PE would be missed/1000 patients screened.73 In practice, therefore, in patients at ‘low risk’ of PE with low D-dimers, a diagnosis of PE can be excluded. These clinical decision rules have not been validated in patients already hospitalised, nor in pregnant women. Hospitalised patients should therefore be investigated by appropriate imaging studies.
Variables used to determine patient pretest probability for pulmonary embolism*
Clinical signs and symptoms of DVT (minimum of leg swelling and pain with palpation of deep veins)
PE as or more likely than an alternative diagnosis
Heart rate > 100
Immobilisation or surgery in the previous 4 weeks
Malignancy (on treatment, treated in the last 6 months or palliative)
*Score > 4, probability of PE is ‘likely’; 4, probability for PE is ‘unlikely’. Alternatively, < 2 is low probability, 2–6 moderate and > 6 high.
Adapted from Wells PS. Integrated strategies for the diagnosis of venous thromboembolism. J Thromb Haemost 2007; 5(Suppl 1):41–50. With permission from John Wiley and Sons.
Investigation For PE
Chest X-ray (CXR)
Radiographic findings in PE are usually non-specific on a CXR; however, it will exclude other diagnoses (such as infection or pulmonary oedema).
Computed tomography pulmonary angiogram (CTPA)
CTPA is now the gold standard for detecting PE, with multislice scanners reporting sensitivities of 83–100% and specificity of 89–97%. These multislice CTPA scanners allow the entire pulmonary arterial tree to be visualised in less than 10 seconds, and in addition to demonstrating the presence or absence of a PE, an alternative cause for the patient's symptoms may be detected (not possible with V/Q scanning – see below). CTPA also gives an assessment of the right ventricular:left ventricular size, an indicator of the severity of the PE in the acute situation. A good-quality negative CTPA on a multidetector scan effectively excludes PE, and anticoagulation can safely be withheld in such patients.74
Ventilation/perfusion (V/Q) scanning
Until recently, V/Q scanning was the imaging investigation of choice for those suspected of having had a PE, but has been largely superseded by CTPA. It is relatively easily performed, is less invasive and cheaper than pulmonary angiography, and can be used as an alternative in patients with contraindications to CTPA. It is of most use in those with normal CXR without underlying lung disease.74 A normal V/Q scan excludes the diagnosis of PE (1% VTE in follow-up).75
Echocardiography can be particularly useful in unstable patients for whom transport to the radiology department is unfeasible. Right ventricular freewall hypokinesis and increased pulmonary pressures are highly suggestive of PE and can be considered diagnostic if ultrasound evidence of DVT is also seen. It may also allow differentiation between other clinical conditions that can present in a similar fashion, such as aortic dissection, myocardial infarction or pericardial tamponade.76
Currently, the clinical utility of MRI in the diagnosis of PE is low when compared with CTPA. Although preliminary studies suggest contrast-enhanced magnetic resonance angiography (ce-MRA) has a similar sensitivity and specificity to other imaging techniques,77 there is less access to patients (who may be clinically unstable) whilst in the magnet, the duration of scanning is longer and MRI scanning is generally less easily available. However, MRI scanning has some advantages: firstly, it does not require ionising radiation, which is beneficial for women of childbearing age and/or those who are pregnant; secondly, MRI contrast agents are considered to have fewer side-effects than the iodinated media required for CT; and thirdly, it is possible to carry out a comprehensive protocol of magnetic resonance venography and pulmonary angiography (taking less than 20 minutes in total). As a result, consideration could be given to using MRI in patients with contraindications to contrast media, young women with a low clinical probability of PE, and in pregnancy. However, the other imaging modalities discussed above remain the mainstay of investigation at the present time.
Investigation in pregnancy
Clinical assessment of DVT and PE is especially unreliable in pregnancy and the minority of pregnant women undergoing objective investigation for VTE will have that diagnosis confirmed. It is important, therefore, that the risks of imaging (both to the mother and the foetus) are minimised. The Royal College of Obstetricians advises that pregnant women suspected of having a PE have a duplex ultrasound performed,78 with V/Q scanning or CTPA reserved for those with negative ultrasound scans.30 CTPA has the disadvantage of a high radiation dose to the woman's breasts and an increased lifetime risk of developing breast cancer, with V/Q resulting in a considerably lower radiation dose to the mother.
Summary Of Diagnostic Methods For PE
Patients presenting with symptoms suggestive of PE should have a clinical assessment and pretest probability determined. Those at low risk with D-dimers below the laboratory ‘cut-off’ for the exclusion of VTE do not usually need further investigation. For those with suspected PE in whom the diagnosis is ‘PE likely’ or ‘PE unlikely but with positive D-dimers’, CTPA should be performed; a negative multislice scan can safely exclude the presence of a PE. Hospital inpatients with symptoms/signs of PE should have imaging investigations performed. Pregnant women should have duplex ultrasound of the leg veins performed as the initial investigation.
Management of venous thromboembolic disease
Aims Of Treatment
The primary aims of the treatment of VTE are to relieve symptoms and to prevent thrombus extension and recurrent thrombotic events, thereby reducing the risks of the long-term complications of VTE (the post-thrombotic syndrome and chronic thromboembolic pulmonary hypertension). Anticoagulation is the mainstay of treatment of VTE for both DVT and PE.
Before starting anticoagulation, assessment should be undertaken to determine whether any disorders predisposing to VTE are present (i.e. malignancy or pregnancy), to assess the safety of proposed anticoagulant therapy and to determine whether appropriate monitoring of anticoagulant therapy is achievable. Thorough clinical history taking and examination will identify factors contributing to the development of VTE and risks for anticoagulant therapy. Baseline blood tests should be undertaken (full blood count, renal function and coagulation screen) to allow safe prescribing of anticoagulant therapy. Heparins are renally excreted, vitamin K antagonists and unfractionated heparin require coagulation test monitoring, and care should be taken with anticoagulant therapy in anaemia and thrombocytopenia.
Anticoagulation should be started as soon as an objective diagnosis of VTE has been made, of if there are any delays in investigation of patients at high risk of VTE.29,72 For most patients, this will mean immediate anticoagulation with an injectable anticoagulant (UFH, LMWH or fondaparinux), followed by oral anticoagulation with a vitamin K antagonist. Rivaroxaban, an oral direct Xa antagonist, is now licensed for the treatment of DVT. Some patients (such as those with PE with haemodynamic compromise or DVT associated with phlegmasia caerulea dolens) may benefit from thrombolytic therapy or even surgical intervention.
Haemodynamically compromised patients with PE should be managed in an appropriate clinical setting – usually a coronary care or high-dependency unit. Haemodyanamic support (inotropes) and oxygen should be administered if required. Intravenous UFH will achieve therapeutic levels faster than LMWH, and the dose can be adjusted if thrombolytic therapy is required; therefore, it should be used in preference to subcutaneous LMWH in shocked patients.
Thrombolysis achieves clot lysis more rapidly than anticoagulation alone, accelerating the reduction in pulmonary vascular obstruction, potentially increasing pulmonary perfusion and gas exchange, and should be considered in patients with haemodynamic instability.72
UFH is renally excreted, has an unpredictable dose response and a narrow therapeutic window; it therefore needs monitoring. It is an injectable anticoagulant and has a short half-life of 60–90 minutes. It is usually administered by a continuous intravenous infusion after an initial loading dose (loading dose 80 U/kg followed by an infusion of 18 U/kg per hour) with subsequent dose adjustments made to maintain an activated partial thromboplastin time (APTT) ratio of 1.5–2.5. Randomised clinical trials have shown that treatment with intravenous (i.v.) UFH for 5–7 days followed by more prolonged treatment with an oral anticoagulant is as efficacious as longer treatment with i.v. UFH.79 Haemorrhage occurs in up to 5% of individuals receiving UFH infusions, the risk being greater in the elderly and those receiving antiplatelet therapies. Care should be taken to monitor for HIT (see earlier). Oral anticoagulation (see below) should be started concurrently with UFH, and the UFH stopped when the INR is therapeutic on two consecutive days.79 Use of UFH has now largely been superseded by LMWH other than in patients with renal failure, or those at high risk of bleeding.
LMWHs are made by chemical or enzymatic depolymerisation of UFH. Their predominant mode of action is to inhibit anti-Xa activity, although all have variable anti-IIa activity. In comparison to UFH, they have more reliable pharmacokinetics and a greater bioavailability and can be given by once or twice daily weight-adjusted subcutaneous administration. A Cochrane review that compared fixed-dose LMWH with UFH for acute PE treatment found an odds ratio of 0.88 (95% confidence interval 0.48–1.63) for risk of recurrent PE in favour of LMWH.80 LMWH does not, routinely, require monitoring. In certain circumstances, however (e.g. in renal impairment or elderly patients with low body weight), monitoring may be of benefit. Peak anti-Xa levels should be measured 3–4 hours after LMWH injection, the therapeutic range being 1.0–2.0 IU/mL with once-daily injection and 0.5–1.0 IU/mL with twice-daily injections. The risk of bleeding is less with LMWH compared with UFH, as is the risk of HIT.
Studies have shown that, for the treatment of DVT, outpatient LMWH administration at therapeutic licensed dosages is as safe and efficacious as UFH infusion.79
LMWHs have been shown to be as safe and effective as UFH in treating non-massive haemodynamically stable PE, but the optimal dosing schedule (once or twice daily) remains controversial. Prognostic prediction models (such as Pulmonary Embolism Severity Index) are being used to categorise risk of mortality, and patients at low risk are suitable for outpatient management or early discharge.81
LMWHs are the treatment of choice for VTE in pregnancy, as warfarin is teratogenic. As the half-life of LMWH in pregnancy is shorter, twice-daily regimens should be utilised for pregnant patients. Furthermore, these patients should be managed in consultation with an obstetrician and a haematologist with an interest in obstetric haematology.78
Studies in cancer patients comparing LMWH to standard warfarin anticoagulation in patients with acute VTE and malignancy demonstrated a significant reduction in the occurrence of recurrent VTE without an increase in bleeding in those receiving LMWH.82 Although there was no reduction in mortality, LMWH was well tolerated and avoided the necessity of frequent INR checks; it therefore should be offered to patients with malignancy and VTE, especially whilst undergoing chemotherapy, which can cause unstable INRs.
Fondaparinux is a pentasaccharide that exerts a selective anti-Xa anticoagulant effect via antithrombin. Once-daily, subcutaneous, fixed-dose fondaparinux (7.5 mg subcutaneously once daily for patients 50–100 kg in weight) has been shown to be as effective and safe as UFH and LMWH for the management of acute symptomatic DVT and PE, and has a licence for these indications. Whilst it is more expensive than UFH and LMWH, it does not need monitoring, nor has it been associated with the development of HIT.83
Vitamin K antagonists
Warfarin remains the treatment of choice for the majority of patients with VTE. It has a narrow therapeutic window, there is considerable variability in dose response between subjects, it has numerous drug and food interactions, and requires regular monitoring. All too frequently there are dosing problems due to patient non-adherence or miscommunication between patient and health professional. Warfarin inhibits the metabolism of vitamin K, which is required for the essential post-translational modification (carboxylation) of the vitamin K-dependent coagulation factors (factors II, VII, IX and X). The rate of inhibition of these coagulation factors is dependent on their rate of synthesis; factor VII (with a half-life of 6 h) is rapidly inhibited, whereas it takes considerably longer for the activity of factor II (with a half-life of 72 h) to fall. It usually takes 4–7 days for the activities of these vitamin K-dependent coagulation factors to fall to a level at which the patient is therapeutically anticoagulated.
Warfarin is monitored by the INR (derived from the prothrombin time). ‘Bridging’ anticoagulation with UFH, LMWH or fondaparinux should be continued for a minimum of 5 days and until the INR is > 2. The target INR for a patient with DVT or PE (unless the thrombosis occurred on warfarin) is 2.5. Not surprisingly, the most serious adverse complication of oral anticoagulation is bleeding. The risk is directly related to the intensity of anticoagulation and the length of therapy, and is greatest in the elderly, those with unstable anticoagulation and those with significant concomitant illnesses (including past history of gastrointestinal bleeding, renal or liver failure, anaemia, uncontrolled hypertension). The risk of major bleeding with warfarin anticoagulation is reported to be 0.5–6.5% per year and fatal bleeding 0.1–1.0% per year.84 For patients newly starting warfarin anticoagulation it is not (yet) possible to predict their dose, and induction algorithms (such as the modified Fennerty algorithm) should be used to guide dosing.85
Rivaroxaban, a once-daily, oral, direct Xa inhibitor has now been licensed in the UK and approved by NICE and the Scottish Medicines Consortium for the treatment of DVT and the secondary prevention of VTE. In the management of acute DVT, rivaroxaban was non-inferior when compared to traditional anticoagulation (LMWH and warfarin) in the prevention of symptomatic, recurrent VTE, without an increase in bleeding.86 Similar findings were seen in the more recently published study in patients presenting with PE,87 but as yet rivaroxaban is not licensed in the acute management of PE.
There are clear advantages to the use of rivaroxaban in the treatment of patients with DVT; it is an oral medication with a reliable pharmacokinetic profile and therefore does not need monitoring. One disadvantage is that there is no direct antidote, although there is some evidence for the use of prothrombin complex concentrates and novoseven in the management of patients bleeding while taking rivaroxaban.88 It is expected that as clinical experience with these new anticoagulants grows, they will, in time, largely replace the use of warfarin.
Thrombolysis has been used for the treatment of DVT, initially given systemically and more recently by local catheter. Thrombolysis has been demonstrated to produce more rapid early clot lysis and reduced incidence of the post-thrombotic syndrome (PTS), but is associated with a significant risk of bleeding. Patients at low risk of bleeding (predominantly young patients), with extensive iliofemoral DVT, may significantly benefit from this treatment.89 Local administration of thrombolytic agents, or the use of combined catheter-directed thrombolysis with mechanical thrombectomy, reduces that bleeding risk.90 Thrombolytic therapy should be considered for patients with recent-onset (within 7–10 days) massive iliofemoral DVT or those with limb-threatening thrombosis.91
Inferior Vena Caval (IVC) Filters
There is no evidence to support the routine use of IVC filters where a patient can be anticoaguated. There is only one randomised trial of the use of vena caval filters in the management of VTE in patients who were anticoagulated; this demonstrated a reduction in symptomatic PE, but a higher risk of recurrent DVT without a reduction in mortality in patients with IVC filters.92
The remainder of the evidence for IVC filter use comes predominantly from descriptive case series and therefore there is little robust evidence to guide the clinician on their use. Both permanent and retrievable filters are available (some may be retrieved up to a few months after insertion). It is recommended that insertion of an IVC filter be considered to prevent PE in patients with VTE and a contraindication to anticoagulation; this includes patients with recent (within 2 months) VTE who require cessation of anticoagulation for surgery. Anticoagulation should be commenced when the contraindication to its use resolves, and should be considered for all patients with IVC filters in situ (dependent on perceived risks of thrombosis from disease and bleeding associated with anticoagulant therapy). It is also reasonable to consider IVC filter insertion in patients who have PE despite therapeutic anticoagulation. If the indication for the filter is temporary, retrievable filters should be used (and then removed) where possible.
Duration Of Anticoagulation
There remains uncertainty about the optimal duration of anticoagulation following a VTE. Systematic reviews have considered the duration of anticoagulation after an episode of VTE and have reported that, while the risk of recurrent VTE is low should anticoagulant therapy be continued, the risk of bleeding is increased.93 Short-term anticoagulation (less than 3 months) is associated with a higher risk of recurrence compared with longer-term treatment.94 Following cessation of oral anticoagulation after a first episode of VTE, the risk of recurrence is 7–12.9% after 1 year and 21.5–22.8% after 5 years.9,95
At least 3 months of anticoagulant therapy is required after a proximal DVT or PE; 3 months is likely sufficient after a first event if it was associated with a transient risk factor, such as surgery. Calf vein thromboses, if diagnosed, should be treated with anticoagulation for 6–12 weeks.85
A prospective study followed 570 patients with a first episode of VTE for 2 years after the cessation of oral anticoagulant therapy.95 The risk of recurrence at 2 years in those who had presented with VTE within 6 weeks of surgery or in pregnancy or postpartum was zero, in contrast to those who had had idiopathic events (19.4%) and those who had had a non-surgical risk factor for VTE (8.8%). Patients with a clear precipitating factor for VTE are at low risk of recurrence once stopping anticoagulation if the underlying risk factor has resolved.
Longer-term anticoagulation after a first idiopathic VTE may be appropriate in patients considered at high risk of recurrent VTE, who are at low risk of bleeding, following an individual assessment of risk factors. As active cancer and anticancer treatment both increase the risk of VTE, consideration should be given to continuing anticoagulation.