Diagnosis and Anticoagulant Treatment *

Diagnosis of DVT

The clinician should maintain clinical vigilance to consider the possibility of deep venous thrombosis (DVT) or pulmonary embolism (PE) that may occur with leg pain or shortness of breath, respectively, but may alternatively have subtle, atypical, or no symptoms. Because the clinical symptoms and signs on their own are unreliable, a suspected DVT should be confirmed by an objective test. Currently, duplex scanning (ultrasonography), which combines venous compression with blood flow and velocity recordings, is the initial investigation of choice.^1–4 The sensitivity and specificity are in excess of 98% for DVT above the knee and in excess of 95% for DVT in the calf.^5–10 One of the advantages for ultrasound is that in the absence of DVT, it can often provide an alternative diagnosis for symptoms such as ruptured Baker cyst or muscle hematoma.

Although performing ultrasonography on every patient suspected of having DVT is feasible, it is expensive and is a strain on ultrasound resources. The combination of a clinical score with a d-dimer assay is an alternative initial approach that can spare many patients from an unnecessary ultrasound examination.

Several clinical scoring systems for DVT have been developed. These are the Wells,^11–13 Khan, ¹⁴ Constans, ¹⁵ and Büller ¹⁶ scoring systems. The Wells scoring system is the one most widely used and it can classify patients into low, moderate, and high pretest probabilities with a prevalence of DVT of 5%, 17%, and 53%, respectively.

d-Dimer enzyme-linked immunosorbent assay is the blood test for suspected DVT or PE. ¹⁷ This is a “rule out” test and VTE is extremely unlikely if the test is normal. However, the d-dimer lacks specificity and will be elevated in acute VTE as well as in multiple other illnesses such as myocardial infarction, cancer, sepsis, the postoperative state, during pregnancy, and following childbirth.

The presence of a normal d-dimer test in patients with a low Wells pretest probability can rule out DVT^11,12 making further investigation with ultrasound unnecessary. It has been demonstrated by studies with a 3-month follow-up that it is safe not to treat such patients with anticoagulants.^3,18–20

Diagnosis of PE

The best diagnostic imaging test for PE is the chest computed tomography (CT) scan. ²¹ Isotope lung scanning has now been relegated to a second-choice imaging test reserved for patients in whom use of contrast agent might be hazardous such as those with renal failure and in order to avoid radiation in young people or the breast. A 16-slice multidetector-row CT, for example, can image the entire chest with a single breath-hold of less than 10 seconds and can identify the entire range of PE from massive saddle embolism to sub-millimeter subsegmental PE in sixth-order pulmonary arterial branches.

For suspected PE, the Wells Scoring System for PE relies upon a weighted point score for 8 items obtained from the history and physical examination and may assist in categorizing clinical likelihood into low, moderate, or high probability with a prevalence of PE of 1.3%, 16.2%, and 37.5%, respectively. ²² Other scores are the simplified Wells score, ²³ the Geneva score, and its simplified version. ²⁴ A multicenter study involving 807 patients demonstrated that any of the 4 scores in combination with a normal d-dimer test showed a similar performance for exclusion of PE. ²⁵ Specificity of the d-dimer test is low (30%-40%) and decreases with age, whereas the risk of PE increases with age, so that it has been suggested that an age-adjusted cutoff point for d-dimer (patient’s age × 10 µg/L) in combination with a low pretest score would increase the number of patients over 50 years in whom PE could be excluded. This approach has been demonstrated in a series of 5132 consecutive patients with suspected PE ²⁶ and has been validated in a series of 414 patients demonstrating that PE could be safely excluded in 19% to 22% of the patients with any of the 4 scoring systems compared to 13% to 14% when the d-dimer cutoff point was not age adjusted. ²⁷

Avoidance of an unnecessary spiral CT scan prevents patients from exposure to substantial ionizing radiation that has significant risks.^28,29 In young nonpregnant women with suspected PE and normal chest x-ray, nuclear perfusion lung scan may be preferred to CT lung scan, because of concern about the degree of lifetime radiation exposure and risk of cancer (eg, breast cancer). In women with suspected or confirmed pregnancy, the mother may likewise prefer nuclear perfusion lung scanning as an alternative to CT lung scanning to reduce fetal radiation exposure. Nuclear ventilation lung scanning is not performed in pregnancy. ³⁰

A meta-analysis involving 2982 patients has indicated that in patients in whom PE has been ruled out by CT pulmonary angiography, the occurrence rate of PE was 1% (95% confidence interval [CI] 0.7%-1.4%) at 3 months. ³¹

General Considerations

The objectives for treating acute DVT are to prevent death and disability from PE, pulmonary hypertension, and peripheral venous disease. Further aims are to prevent recurrence of VTE and development of postthrombotic syndrome (PTS) as a result of persistent venous outflow obstruction and/or dysfunction of the venous valves. ³² Extension of recurrent DVT into the collateral circulation produces further outflow obstruction and progressive swelling of the leg. Massive extension can result in increased compartmental pressure possibly leading to phlegmasia cerulea dolens, which although rare and often associated with metastatic cancer can lead to venous gangrene and limb loss.

It has been demonstrated that asymptomatic below-knee DVT can lead to subsequent development of the PTS^33,34 and that 18% of the symptomatic calf DVT are associated with proximal extension or recurrence ³⁵ indicating that below-knee DVT merits treatment.

Anticoagulants

In patients with DVT, initial therapy with vitamin K antagonist (VKA) alone is associated with an unacceptable high rate of recurrent symptomatic VTE. ³⁶ Also, extension of DVT was observed in 39.6% of the patients on VKA alone, but only in 8.2% of the patients treated initially with heparin and subsequently with VKA (P < .001). ⁶ Thus, initial parenteral heparin and subsequent long-term oral anticoagulation are both necessary.^32,36

Findings from randomized clinical trials (RCTs) in the 1990s resulted in low-molecular-weight heparin (LMWH) replacing unfractionated heparin (UFH) in the initial treatment of DVT. These studies concluded that LMWH is at least as effective and safe as initial treatment for acute VTE compared to intravenous UFH.^37–47 The LMWH was also found to be as effective and safe as intravenous UFH in patients with acute PE.^48–50 Thus, anticoagulation should usually be started with LMWH for patients with PE. Treatment with intravenous UFH, which generally requires hospitalization, is now less frequently used but remains preferable therapy in patients with massive or submassive PE in the presence of chronic kidney disease in view of the increased risk of bleeding in such patients.

Several studies suggested that when using UFH for the initial treatment of DVT, rapid achievement of an activated partial thromboplastin time within the therapeutic range (2.0-3.0 times the control) within 24 hours reduces the rate of recurrent DVT.^51–53 However, other studies did not confirm this finding.^54,55

In contrast to UFH, LMWHs have a consistent dose–response with predictable bioavailability when given subcutaneously (sc). They do not require hematologic monitoring apart from the platelet count. The need for anti-Xa monitoring has been reduced by specific labeling of individual regimens in the context of renal insufficiency or obesity (see pharmacopeia). They may be administered once a day.^40,56–59 These properties have made LMWH the preferred treatment for patients with uncomplicated DVT as outpatients.^32,60–68 The LMWH should be administered for at least 5 days^69,70 and should be discontinued when the patient’s international normalized ratio (INR) is stable within the therapeutic range of 2.0 to 3.0.

The RCTs have demonstrated that fondaparinux is as effective as intravenous UFH for the initial treatment of DVT and PE.^71,72 Fondaparinux is administered once daily. Heparin-induced thrombocytopenia is rare. Attention to labeling is essential in patients with impaired renal function in whom the risk of bleeding is increased.

The VKA treatment should be adjusted to maintain the INR between 2.0 and 3.0 (target INR 2.5). The risk of bleeding in relation to different INR ranges as reported by several studies is shown in Table 14.1.^73–75 An INR greater than 4.0 is associated with an increased frequency of hemorrhagic complications.^76–78 The VKA may be started on the first day of heparin therapy, except when patients require thrombolysis or surgery or where there are comorbidities that predispose to major bleeding.^69,70,79 Whether low-dose warfarin, which produces a targeted INR between 1.5 and 1.9, may offer a suitable option for patients requiring extended periods of anticoagulation has long been debated. One study showed a definite advantage for low-dose warfarin over placebo in patients who had completed an initial 6.5-month period of conventional anticoagulation when compared to placebo. In this study, 508 patients who had been on full VKA therapy for 6.5 months were randomized to low-intensity warfarin or placebo. There were 37 recurrences in the placebo group of 253 patients (7.2 per 100 person years) and 14 in the low-intensity warfarin group of 255 patients (relative risk [RR] 0.36; 95% CI 0.19-0.67; P = .001). ⁸⁰ However, in a direct comparison between conventional anticoagulation and low-dose warfarin, the former proved to be more effective and equally safe. ⁷⁵ Indeed, in this study involving 738 patients with unprovoked proximal DVT or PE, the incidence of recurrence over a 2 to 4 years of follow-up increased from 2% in the conventional intensity treatment group to 4% in the low-intensity treatment group (RR 2.67; 95% CI 1.05-6.74). The incidence of major hemorrhage was 2% in each group. Thus, the risk of recurrent VTE increases even with INR of less than 2. We believe that conventional warfarin regimen should be regarded as the first choice. However, a low-intensity regimen can be considered in particular situations depending on individual judgment, for example in patients reputed to be at a higher hemorrhagic risk and in those who have a strong preference for less frequent INR monitoring.

Table 14.1.

Major Bleeding Complication Rate According to INR Intensity.

Trial	INR Range	Event Rate Per 100 Person-Years
Kearon et al 73	2.0-3.0	3.8
Schulman et al 74	2.0-2.85	2.4
Kearon et al 75	2.0-3.0	0.9
Kearon et al 75	1.5-1.9	1.1

Abbreviation: INR, international normalized ratio.

Rivaroxaban is a new oral direct inhibitor of Xa. In a phase III noninferiority study, 3449 patients with acute, symptomatic DVT were randomized to rivaroxaban (15 mg twice daily for 3 weeks, followed by 20 mg once daily without initial parenteral therapy) or sc enoxaparin followed by VKA for 3, 6, or 12 months (duration according to treating physician’s discretion). Recurrent VTE occurred in 2.1% in the rivaroxaban group and in 3.0% in the control group (RR 0.70; 95% CI 0.46-1.07; P < .0001 for noninferiority and P = .076 for superiority of rivaroxaban). Major bleeding or clinically relevant nonmajor bleeding occurred in 8.1% of the patients in each group. ⁸¹

The efficacy of rivaroxaban in the prevention of recurrent VTE was tested in the EINSTEIN-extension study performed in parallel and reported in the same publication. In this study, 1197 patients who had completed their anticoagulation (6-12 months) were randomized to continue with rivaroxaban or placebo for a further 6- to 12-month period. The recurrence rates for VTE were 1.3% in the rivaroxaban group and 7.1% in the placebo group (RR 0.22; 95% CI 0.11-0.45; P < .001). The nonfatal major bleeding rate was 0.7% in the rivaroxaban group and zero in the placebo group (P = .11). ⁸¹

In a RCT involving 4832 patients who had symptomatic PE with or without DVT, rivaroxaban (15 mg twice daily for 3 weeks, followed by 20 mg once daily) was compared to standard therapy (enoxaparin followed by an adjusted dose of VKA) for 3, 6, or 12 months. Rivaroxaban was noninferior to standard therapy for symptomatic recurrent PE (RR 1.12; 95% CI 0.75-1.68; P = .003 for noninferiority). Major bleeding was 1.1% in the rivaroxaban group and 2.2% in the standard-therapy group (RR 0.49; 95% CI 0.31-0.79; P = .003). ⁸²

Apixaban, an oral reversible inhibitor of factor Xa, was tested in a dose-ranging study involving 520 consecutive patients with symptomatic DVT against standard therapy (LMWH for a minimum of 5 days followed by VKA) for 3 months. ⁸³ Symptomatic recurrence of VTE and extension of thrombus as detected by ultrasound occurred in 4.7% of the patients in the apixaban groups (it was comparable in all 3 groups) and 4.2% in the standard therapy group. The rate of major and clinically relevant nonmajor bleeding was 7.3% in the apixaban groups and 7.9% in the standard therapy group. Phase III studies are in progress.

Dabigatran is a new oral direct inhibitor of thrombin. In a phase III noninferiority study, 2539 patients with acute symptomatic DVT who were initially given parenteral anticoagulation therapy for 8 to 11 days were randomized to dabigatran or sc heparin (UFH or LMWH) followed by VKA for 6 months. Recurrent VTE occurred in 2.4% in the dabigatran group and in 2.1% in the control group (RR 1.10; 95% CI 0.66-1.84; P < .001 for noninferiority). Major bleeding occurred in 1.6% of the patients in the dabigatran group and in 1.9% in the standard therapy group (RR 0.83; 95% CI 0.46-1.49). Adverse events leading to discontinuation of the study drug occurred in 9.0% of patients in the dabigatran group and in 6.8% of patients in the warfarin group (P = .05). ⁸⁴

The efficacy of dabigatran in the prevention of recurrent VTE was tested in 2 subsequent studies. In the first (RE-SONATE study) 1343 patients who had completed their anticoagulation (6-18 months) therapy were randomized to continue with dabigatran or placebo for a further a 6-month period. ⁸⁵ The recurrence rate for VTE was 0.4% in the dabigatran group and 5.6% in the placebo group (RR 0.08; 95% CI 0.02-0.25; P < .001). Nonfatal major bleeding occurred in 0.3% of the dabigatran group and zero in the placebo group (P = .996). In the second (RE-MEDY study) 2856 patients who had completed their anticoagulation (3-12 months) therapy were randomized to receive dabigatran or conventional warfarin for up to 36 months. ⁸⁶ The recurrence rate for VTE was 1.8% in the dabigatran group and 1.3% in the warfarin group (RR 1.44; 95% CI 0.78-2.64; P < .027 for noninferiority). The rate of major bleeding was 0.9% in the dabigatran group and 1.8% in the warfarin group (hazard ratio 0.52; 95% CI 0.27-1.02; P = .058). In this study, a higher number of acute coronary syndromes were observed during treatment with dabigatran compared to warfarin (0.9% vs 0.2%; P = .02).

The efficacy of aspirin (100 mg daily for 2 years) in the prevention of recurrent VTE was recently investigated in a RCT involving 402 patients who had completed 6 to 18 months standard therapy for first-ever unprovoked VTE. The incidence of recurrent VTE was 6.6% in the aspirin group and 11.2% in the placebo group (RR 0.58; 95% CI 0.36-0.93). One patient in each group had major bleeding. ⁸⁷ Thus, extended treatment with aspirin may be an appropriate choice in patients who are at high risk of bleeding with VKA. However, confirmatory studies are needed. It should be noted that the 42% reduction of recurrent VTE reported in the above study is approximately half of that produced by rivaroxaban and dabigatran. Compared to placebo, these oral anticoagulants reduced the risk of recurrent VTE by more than 80% (see above).

Long-Term Treatment With LMWH

Five studies involving 1818 patients compared the effect of therapeutic or near-therapeutic LMWH doses for 3 to 6 months on VTE recurrence compared to conventional VKA therapy,^88–92 mainly in noncancer patients although 3 studies included some patients with cancer.^88,91,92 One reported the results in the patients with cancer separately. ⁹¹ The incidence of recurrent VTE was 4.0% in the LMWH groups and 6.2% in the VKA groups (RR 0.68; 95% CI 0.45-1.022).

Four studies involving 1201 patients compared the effect of therapeutic or near-therapeutic LMWH doses for 3 to 12 months on VTE recurrence compared to conventional VKA therapy in patients with cancer.^93–96 The number of patients involved was 1201 including the patients with cancer from the study above that reported the results in the patients with cancer separately. The incidence of recurrent VTE was 7.5% in the LMWH groups and 16.1% in the VKA groups (RR 0.46; 95% CI 0.33-0.65).

The incidence of major bleeding in all the studies reported above involving patients without cancer and with cancer was 3.2% in the LMWH group and 3.9% in the VKA group (RR 0.83; 95% CI 0.56-1.22). ⁹⁷

It appears that long-term LMWH is equally effective as standard therapy for preventing recurrent VTE in patients without cancer, but more effective for patients with cancer.

Standard treatment of DVT (initial LMWH for 5 days followed by VKA) prevents thrombus extension and embolization but does not directly lyse the thrombus, and this frequently results in partial recanalization. A number of studies that compared long-term treatment with LMWH versus standard therapy demonstrated better recanalization in the long-term LMWH groups.^{90,91,98–101} A meta-analysis on 5 studies that reported on total recanalization demonstrated a RR of 0.66 (95% CI 0.57-0.77; P < .0001) in favor of long-term LMWH. ¹⁰² In a large multicenter study involving 480 patients, there was a reduction in PTS (RR 0.77; P = .001). ⁸⁹ Pooled analysis on 2 studies reporting on the subsequent development of leg ulcers^89,103 yielded an 87% risk reduction for venous ulcers with LMWH (P = .019). ¹⁰²

Idraparinux is a synthetic pentasaccharide that inhibits factor Xa mediated through antithrombin. In a RCT involving 2904 patients with DVT, treatment with idraparinux (2.5 mg sc once weekly) was associated with DVT recurrence in 2.9% compared to 3% in the standard therapy group (initial heparin followed by VKA) at 3 months. These results satisfied the prespecified noninferiority requirement. Clinically relevant bleeding was 4.5% in the idraparinux group and 7% in the standard therapy group (P = .004). Bleeding rates were similar at 6 months. ¹⁰⁴

A second RCT involving 1215 patients who had completed 6 months of treatment with an anticoagulant (idraparinux or VKA), compared 2.5 mg of sc idraparinux weekly with a placebo for a further 6 months. The incidence of recurrent VTE was 1% in the idraparinux group and 3.7% in the placebo group (P < .001). Patients on idraparinux had a higher incidence of major bleeding (3.1% vs 0.9%). ¹⁰⁴

Two studies investigated the efficacy and safety of idraparinux in the treatment of PE. In the first study, which involved 2,215 patients, the incidence of recurrence in the idraparinux (2.5 mg sc weekly) at 3 months was 3.4% compared to 1.6% in the standard therapy group (odds ratio [OR] 2.14; 95% CI 1.21-3.78). These results did not meet the noninferiority requirement. ¹⁰⁴ In the second study, 3202 patients with PE were randomized to 5 to 10 days of enoxaparin followed by idrabiotaparinux 3.0 mg weekly or warfarin (INR 2.0-3.0) for 3 or 6 months. Idrabiotaparinux has the same pharmacodynamic effects as idraparinux but has the advantage of rapid neutralization by intravenous avidin. The incidence of recurrent PE was 2% in the idrabiotaparinux group and 3% in the warfarin group (P for noninferiority = .0001). Clinically relevant bleeding occurred in 5% of the patients in the idrabiotaparinux group and 7% in the warfarin group (OR 0.67; 95% CI 0.49-0.91; P for superiority = .0098). It is of interest that the efficacy of idrabiotaparinux given for 3 to 6 months persisted beyond the end of treatment until at least 1 year, whereas there was an almost immediate consistent increase over time for recurrent VTE in the control population after ceasing warfarin. ¹⁰⁵

Duration of Anticoagulation Therapy

The aim of extending the duration of treatment is to prevent recurrent DVT that depends on several risk factors. The risk is low if DVT occurs in the presence of a reversible risk factor, but the risk is high if DVT is unprovoked^106–117 or occurs in the presence of active cancer.^{106,112,113,118} Patients with symptomatic PE have a higher risk of PE recurrence than those with DVT alone. ¹¹⁹

The lowest risk is found when surgery is the reversible risk factor.^109,120 The estimated 5-year cumulative risk of recurrent VTE after stopping anticoagulation is 3% if proximal DVT is provoked by surgery, 15% if provoked by a nonsurgical reversible risk factor, and 30% if unprovoked. ⁹⁷ The RR is 2.0 for proximal DVT or PE compared to calf DVT,^{111,114,116,117,119} 1.5 if DVT is a second episode,^93,118,121 2.0 if antiphospholipid antibody is present,^73,122–124 1.5 in the presence of a hereditary thrombophilia,^{73,80,109,110,124–132} 1.5 in the presence of residual thrombosis in the proximal veins,^{73,124,133–138} and 1.6 for male gender.^139,140 The risk is higher in the presence of multiple risk factors, homozygous-inherited thrombophilia, or a combination of heterozygous thrombophilias (see section on thrombophilia).

In patients with unprovoked VTE, almost every contemporary trial has found that prolonged anticoagulation with VKA reduces long-term recurrence by about two-thirds ¹¹⁷ but increases the risk of major bleeding. ¹⁴¹ In addition, while the case-fatality rate of major bleeding complications is consistently around 11% that for recurrent VTE decreases after completing an initial treatment period of 3 to 6 months from 11% to 3.6%. ¹⁴² Accordingly, the benefit-to-risk for indefinitely prolonging anticoagulation in patients with unprovoked VTE should be carefully assessed and individually tailored.

Unfortunately, there is no validated prediction tool to stratify the risk of major bleeding during extended anticoagulant therapy for patients with VTE, but this risk appears to increase with the prevalence of the following factors: age >65 years, additional increase in risk if age >75 years, previous gastrointestinal bleeding if not associated with a reversible cause, previous noncardioembolic stroke, chronic renal impairment, chronic hepatic impairment, anemia, thrombocytopenia, diabetes, concomitant antiplatelet therapy, other chronic illness that impairs platelet function, poor control of anticoagulation (ie, variable INR results) including suboptimal access to monitoring or active cancer with additional increased risk if there is metastatic cancer. ¹⁴³

The VKA generally should be continued for a minimum of 3 months.^{107,116,117,135} Three months is sufficient for patients with a reversible risk factor. For patients with a known irreversible major risk (protein C, protein S, lupus anticoagulant) or without a known risk factor, prolonged anticoagulant therapy with periodic reassessment should be considered.^32,116 The review process involves balance of benefit and harm. Patients presenting with recurrent DVT should be treated with a more prolonged anticoagulation regimen compared to those having a first episode. ⁷⁴ The optimal duration of oral anticoagulant therapy depends on the risk of VTE recurrence.

Isolated calf DVT

A randomized study of 51 patients with isolated calf DVT, of whom 23 received warfarin for 3 months and 28 did not, investigated the rate of recurrence. ¹⁴⁴ Recurrences and their extent were confirmed with venography. Both groups received an initial course of heparin and all wore compression stockings. During the first 3 months, recurrence occurred in 29% of the patients in the non-warfarin group compared to none in the warfarin group (P < .01). Five of these patients had recurrence with proximal extension and 1 had a pulmonary embolus. At 1 year, 1 of 23 patients in the warfarin group had a recurrence, compared to 19 of 28 in the non-warfarin group (RR 0.13; 95% CI 0.02-0.99). The findings indicate that oral anticoagulants should be given to all patients with symptomatic isolated calf DVT and that 3 months seem to be sufficient.

Four to 6 weeks versus 3 to 6 months

Four studies involving 1988 patients with a first unprovoked DVT (mainly proximal) or PE compared 4 to 6 weeks anticoagulation with VKA with 3 or 6 months. Follow-up was 1 to 2 years.^{107,108,116,135} The incidence of recurrence was reduced from 12.6% in the 4 to 6 weeks group to 6.7% in the 3 to 6 months group (RR 0.53; 95% CI 0.40-0.71). The incidence of major hemorrhage was increased from 0.61% in the 4 to 6 weeks group to 1.0% in the 3 to 6 months group (RR 1.65; 95% CI 0.60-4.53).

Three months versus 6 to 12 months

Four studies involving 1736 patients with first unprovoked DVT (mainly proximal) or PE compared 3 months of anticoagulation with VKA with 6 or 12 months.^{114,134,145,146} Follow-up was 1 to 3 years. The incidence of recurrence was 9.7% in the 3-month group and 9.6% in the 6- to 12-month group (RR 0.99; 95% CI 0.74-1.32). The incidence of major hemorrhage was increased from 0.93% in the 3-month group to 2.4% in the 6 to 12-month group (RR 2.5; 95% CI 1.16-5.83).

Three to 6 months versus indefinite anticoagulation

Four studies involving 676 patients, the majority with second unprovoked DVT (mainly proximal) or PE, compared 3 to 6 months of anticoagulation with VKA (INR 2-3) with indefinite duration of anticoagulation.^{73,74,147,148} Follow-up was 1.4 to 4 years. The incidence of recurrence was reduced from 18.8% in the 3- to 6-month group to 2.7% in the indefinite duration group (RR 0.18; 95% CI 0.09-10.36). The incidence of major hemorrhage was increased from 1.5% in the 3-to 6-month group to 4.6% in the indefinite duration group (RR 3.03; 95% CI 1.12-8.19).

d-Dimer as a Guide to Continue Anticoagulation

Elevation of d-dimer has been reported to increase the risk of recurrent VTE. A meta-analysis of 4 studies which included 1539 patients indicated a 16.6% recurrence of DVT in patients with elevated d-dimer 1 month after discontinuation of VKA therapy compared to a 7.2% rate in those with normal levels of d-dimer (RR 2.30; 95% CI 1.71-3.10). ¹⁴⁹ A meta-analysis of 7 prospective studies involving 1818 patients investigated the association between elevated d-dimer and VTE recurrence in patients with a first unprovoked VTE episode (DVT, PE, or both). In a Cox proportional hazards model, a positive d-dimer 1 month after cessation of anticoagulation had a hazard ratio of 2.59 (95% CI 1.90-3.52). Of all the other factors studied which included inherited thrombophilia, only male sex had a significant effect on risk. ¹²⁶ Thus, an elevated d-dimer is an indication to continue anticoagulation therapy. However, cessation of antigoagulation does not imply absence of recurrence.

Residual Thrombosis as a Risk Factor for Recurrence

A systematic review (11 studies; 3203 patients) showed a positive relationship between residual thrombosis and recurrent VTE during follow-up. ¹⁵⁰ A subsequent systematic review and meta-analysis of 9 prospective cohort studies and 5 RCT demonstrated that the presence of residual venous obstruction was not associated with increased risk of recurrent VTE (OR 1.24; 95% CI 0.90-1.7) in patients with unprovoked DVT who stopped oral anticoagulation therapy. However, residual venous obstruction was associated with recurrent VTE in patients with any (unprovoked or provoked) DVT (OR 1.5; 95% CI 1.1-2.0). ¹³⁷ In a recent randomized trial, recurrent VTE developed in 17.2% of the patients allocated to conventional fixed anticoagulant duration (3 months for provoked DVT and 6 months for unprovoked DVT) and in 11.9% of those randomized to flexible duration according to persistence of residual vein thrombosis, leading to an adjusted RR of 0.64 (95% CI 0.39-0.99). ¹⁵¹ More studies on this flexible approach are needed.

Compression Therapy and the Post-Thrombotic Syndrome

Effective compression reduces edema and minimizes damage in the microcirculation.^152,153 Four RCTs involving 745 patients demonstrated that elastic compression for 2 years in patients with proximal DVT reduced the incidence of PTS from 39% to 19% (RR 0.49; 95% CI 0.38-0.62).^154–157 It appears that treatment with LMWH combined with early ambulation and elastic compression will further prevent the PTS.^158,159

Recommendations for Treating VTE

Methods of Treatment

Initial treatment is with intravenous UFH, LMWH, or fondaparinux for at least 5 days (level of evidence: high). The LMWH is preferred in most patients. The VKA therapy should be commenced on day 1 and continued according to the INR. Initial therapy with LMWH, intravenous UFH, or fondaparinux should be discontinued when the stable INR is in the therapeutic range (2.0-3.0; level of evidence: high).

Rivaroxaban or dabigatran are an alternative therapy in countries where they have been approved (level of evidence: high). Although the former can be used as a single therapy, the latter should be preceded by 1 week of parenteral anticoagulation with either LMWH or fondaparinux. In patients with a history of cancer, LMWH for 3 to 6 months is the initial treatment (level of evidence: high; see section on cancer for evidence). During pregnancy, LMWH is the treatment of choice throughout pregnancy and for the first 6 weeks after delivery (level of evidence: low; see section on pregnancy for evidence). The LMWH for 3 to 6 months is an alternative to VKA therapy (level of evidence: high). Isolated calf DVT should be treated for 3 months (level of evidence: moderate) or followed by serial ultrasonography on 2 occasions if anticoagulation is contraindicated (level of evidence: low).

Duration of Anticoagulation Therapy

All patients should receive long-term antithrombotic therapy for at least 3 months (level of evidence: high). In patients with a major provoking risk factor that has been removed 3 months is sufficient (level of evidence: high).

In patients with an unknown risk factor, the duration of anticoagulant therapy may be indefinite (level of evidence: high). The decision as to the length of therapy is based upon the balance of benefit and harm/bleeding and the patient’s preference. Patients on continued therapy should undergo periodic reconsideration (level of evidence: low). The review process involves balance of benefit and harm. In patients at lower risk of bleeding and continuing with VKA treatment, patient preferences are considered.

In patients with a minor provoking risk factor, the duration of anticoagulant therapy is uncertain and should be based once again upon the same principles (level of evidence: low).

Role of d-dimer: the evidence to date suggests a role for the use of the d-dimer test for establishing the duration of therapy but current data are insufficient.

The concept of residual DVT by imaging suggests a role for the use of follow-up imaging to establish the duration of long-term antithrombotic therapy but current data are insufficient.

In patients with more than 1 episode of VTE, the duration of anticoagulant therapy is indefinite (level of evidence: high). For long-term prevention of recurrent VTE in patients requiring indefinite anticoagulation rivaroxaban or dabigatran can be considered after completing 3 to 12 months of conventional anticoagulation (level of evidence: moderate) when approved.

Immediate mobilization with graduated elastic compression stockings to be worn for at least 2 years (level of evidence: high) at an ankle pressure of 30 to 40 mm Hg (class II) leads to a more rapid reduction of pain and swelling and reduces the occurrence of PTS.^{154,155,158–160}

The LMWH and Renal Insufficiency

Prophylactic doses

An increased risk of bleeding has not been reported in patients with renal insufficiency receiving prophylactic dosages of LMWH. However, it is advised that for prophylaxis in patients with severe renal insufficiency, prophylactic doses of LMWH should be adjusted down according to creatinine clearance and manufacturer’s instructions.

Therapeutic doses

Enoxaparin (see label), fondaparinux (see label), dalteparin, and tinzaparin may have problems in severe renal failure because they are eliminated mainly through the kidneys.

In patients with renal insufficiency, LMWH in therapeutic doses poses a high risk of major bleeding due to its prolonged half-life. The actual risk of major bleeding has not been assessed in prospective studies. Such studies would have to be done with each LMWH because of different pharmacological properties. Major bleeding in patients with a creatinine greater than 2 mg/dL and a similar number of patients receiving enoxaparin at equal or greater doses for the same indications has been assessed in one retrospective study. Major bleeding occurred in 1 (2%) of 50 patients with normal renal function and 16 (30%) of 53 patients with serum creatinine greater than 2 mg/dL (P < .001). ¹⁶¹

Although protamine sulfate is efficacious in stopping LMWH-induced bleeding in some animal models, there are only limited data for humans.