Medical anticoagulation consists of 3 periods: acute (the first 5–21 days), long-term (3–6 mo), and extended period (thereafter; Fig. 1). In the acute period, anticoagulants are loaded via parenteral heparin lead-in or a higher dose of oral anticoagulants is administered. For long-term treatment, the general anticoagulant treatment period for postoperative VTE is 3 months because of the low risk of recurrence, but is prolonged to 6 months for unprovoked VTE. Extended treatment can be considered in some cases with a high risk of VTE recurrence.

For the treatment of VTE, heparin and VKA have been used widely for a long time, although DOACs are quickly replacing them because of their convenience of use. However, in certain situations or for patients who have already been using older anticoagulants for a long time, heparin and VKA can continue to be used [7]. LMWH is also frequently used, especially in cases of CAT. Compared to unfractionated heparin (UFH), it is more likely to show a predictable dose-response relationship. It has a longer half-life, allowing it to be administered subcutaneously once a day or twice a day, with a lower risk of heparin-induced thrombocytopenia and osteoporosis. Because LMWH is excreted mainly by the kidney, it should not be used in patients with severe renal impairment. Conversely, UFH has a shorter half-life and can thus be useful in patients who require an immediate anticoagulant effect to support rapid conversion. VKA is absorbed in the gastrointestinal tract and metabolized in the liver. It exhibits moderate anticoagulant effects by inhibiting vitamin K between 48 and 72 h after the start of administration. Because protein C and protein S, which are considered natural anticoagulants, may also decrease during the priming of VKA, the risk of thrombosis and skin necrosis increases in the first 2 or 3 days of VKA use. Therefore, during this period, VKA should be administered with heparin. The effect of VKA is affected by the patient’s liver function, concurrent medication, and food intake; therefore, careful monitoring is required.

Dabigatran directly inhibits factor IIa, while rivaroxaban, apixaban, and edoxaban inhibit factor Xa. Several large-scale phase 3 studies have compared DOAC therapies with VKA, confirming that DOACs are not inferior in terms of antithrombotic efficacy and bleeding [8-11]. Thus, DOACs have been used instead of VKA for the treatment of VTE. The major drug absorption sites of rivaroxaban, dabigatran, and edoxaban are the stomach and the proximal small intestine. Therefore, in patients with previous total gastrectomy, these drugs should be taken with caution because they may not be absorbed efficiently. Unlike other DOACs, apixaban is mostly absorbed in the distal small intestine and ascending colon and can be safely used in patients who have undergone gastrectomy; however, it should be avoided in patients who have previously undergone resection of the ascending colon [12]. Dabigatran is produced as a capsule formulation. Thus, when the capsule of dabigatran is opened and the contents are taken alone, its bioavailability rises rapidly. Opening, crushing, or chewing dabigatran capsules is prohibited.

For the use of DOACs in cases of CAT, see the section ‘Treatment of cancer-associated venous thromboembolism’.

Anticoagulation in renal insufficiency: Because the main excretion site of DOACs is the kidney, the anticoagulant effect of DOACs increases in the presence of renal insufficiency, and the likelihood of bleeding complications also increases. Since patients with renal insufficiency were not included in the clinical trials of DOACs (for example, a creatinine clearance of <30 mL/min was an exclusion criterion for an edoxaban trial [13]), attention should be given to the use of DOACs in cases of renal insufficiency. Conversely, LMWH, such as dalteparin, can be used safely in patients with moderate renal impairment. The use of anticoagulants according to renal function is summarized in Table 1.

Anticoagulation in thrombocytopenia: In cases of CAT with thrombocytopenia, mostly in hematologic malignancies, LMWH is the preferred anticoagulant because of its shorter half-life and better safety profile than VKA. DOACs may also be considered for these patients. However, the pivotal trials of DOACs for CAT excluded patients with a platelet count <50,000×10⁹/L [13], or even <75,000×10⁹/L [14]; thus, there is insufficient data on whether DOACs can replace LMWH in the setting of CAT combined with significant thrombocytopenia. Therefore, the use of DOACs in patients with thrombocytopenia requires caution, and decisions should be individualized until more robust evidence is gathered.

Thrombolytic therapy and inferior vena cava filter insertion: Sometimes, thrombolytic therapy can be administered either through systemic infusion via a vein or directly to the thrombus by inserting a catheter. However, as the risk of bleeding increases, this method should be considered only in cases of hemodynamically unstable PE. An inferior vena cava (IVC) filter can be considered in patients with DVT to prevent PE when anticoagulation is contraindicated (e.g., bleeding), but the general use of IVC filters or permanent indwelling hardware should be avoided [15].

Treatment of proximal deep vein thrombosis: Proximal lower extremity DVT occurs when a thrombus is located in the popliteal, femoral, or iliac veins. Clinical symptoms vary according to the anatomy, extent, and degree of occlusion, and range from asymptomatic to extensive edema and gangrene. Anticoagulant therapy is indicated for all patients with proximal DVT [16]. In the case of active hemorrhage, lower platelet counts (<50,000×10⁹/L) or prior intracerebral hemorrhage, an IVC filter is preferable. Recommendations for the treatment of provoked proximal DVT includes 3 months of anticoagulant therapy [16]. In patients with an unprovoked proximal DVT, extended anticoagulant therapy (at least 3 months and potentially indefinite) is preferred, while in patients with high bleeding risk, 3 months of anticoagulant therapy is recommended [16]. Therapeutic options for proximal DVT include VKA, LMWH or DOACs. A selection among these agents is usually made based on clinician experience, as well as the risks of bleeding, patient comorbidities, preferences, cost, and convenience. Thrombolytic therapy is not recommended, except for patients with massive iliofemoral or femoral DVT with a high risk of limb gangrene [16].

Treatment of isolated distal deep vein thrombosis: Isolated distal DVT is diagnosed when a thrombus forms below the popliteal vein without extension to the proximal veins. The decision of whether to treat isolated DVT is controversial, as the 3-month recurrent VTE risks of isolated distal DVT between the untreated and treated groups were discordant between studies [17-20], although the absolute risk of VTE recurrence is reportedly lower than that of proximal DVT or PE. In addition, in contrast to proximal DVT and PE, a guideline for the management of isolated DVT is not supported by high-level evidence. A meta-analysis of 24 studies involving 4,072 patients reported that anticoagulation reduced VTE recurrence rate compared to no anticoagulant treatment [odds ratio (OR), 0.50; 95% confidence interval (CI), 0.31–0.79], without increasing the rate of major bleeding (OR, 0.64; 95% CI, 0.15–2.73) [21]. In contrast, based on the evidence of a randomized placebo-controlled prospective trial [22], asymptomatic patients without high-risk features of VTE recurrence or extension (immobilization, active cancer, unprovoked VTE) and with a high risk of bleeding, may undergo surveillance with serial screening for DVT of the lower leg without anticoagulation. However, if patients are acutely symptomatic, have high-risk features for recurrence and extension, and are without bleeding risk, anticoagulation is recommended [16]. Aspirin is often used instead of an anticoagulant, considering the less aggressive nature of isolated distal DVT, but this method lacks firm evidence [23]. After the decision to anticoagulate, the choice of drugs and duration may follow the treatment strategies for proximal DVT of the lower extremities.

Treatment of hemodynamically unstable pulmonary embolism: Patients with life-threatening PE may require additional treatment beyond anticoagulation, including systemic thrombolysis, catheter-directed therapy, and embolectomy. A meta-analysis showed that the use of thrombolytics was associated with lower all-cause mortality. The reduction in mortality seems to be mainly accounted for by studies including hemodynamically unstable patients [24]. Anticoagulation should be started when the patient becomes stable, without bleeding complications after thrombolysis.

Treatment of hemodynamically stable pulmonary embolism: Anticoagulation is the mainstay of treatment for patients with confirmed PE and should be given to all patients with suspicion of the disease in the absence of active bleeding, even before a diagnosis of PE is confirmed. The anticoagulation of patients with PE has changed considerably in recent years, since DOACs were introduced. DOACs have shown similar efficacy and safety to those of conventional anticoagulants in patients with acute PE in several large randomized clinical trials [8-11, 25]. The risk reduction for recurrent VTE with different DOACs has not been directly compared, but based on indirect comparisons, it appears to be similar among all DOACs.

The recommended treatment duration for PE is usually 6 months but may range from 3 months in patients with a transient risk factor to indefinite in patients with ongoing major risk factors (e.g., cancer, recurrent unprovoked PE).

Treatment of subsegmental pulmonary embolism: Subsegmental pulmonary embolism (SSPE) is diagnosed when PE does not involve the proximal pulmonary arteries. The decision on whether to treat asymptomatic SSPE patients remains to be an issue; a recommendation cannot be supported by high-quality randomized clinical trials, because computed tomographic (CT) angiography may result in a false-positive diagnosis (especially in cases with a single subsegmental involvement in 1 image and normal D-dimer), and the data with respect to whether SSPE could be progressive or recurrent without anticoagulation are still controversial [26, 27]. Patients with SSPE should undergo bilateral Doppler ultrasonography of the lower extremities [28], and high-risk DVT sites, such as upper extremities with indwelling central venous catheters. Patients who are asymptomatic, have no co-existing DVT or high-risk features [29] of recurrent or progressive VTE (immobilization, active cancer, unprovoked VTE), and are at high risk of bleeding may undergo surveillance with serial screening for DVT of the lower leg without anticoagulation. However, if patients are symptomatic, have co-existing DVT and/or high-risk features, and have no bleeding risk, anticoagulation should be suggested rather than surveillance only. When deciding on anticoagulation, the choice of drugs and duration may follow the treatment strategies recommended for PE in larger pulmonary arteries.

Incidental pulmonary embolism: Incidental PE is frequently reported on enhanced chest CT scans. Although the embolic load in incidental PE is usually lower than that in symptomatic PE, anticoagulation should be initiated if it involves a location proximal to the subsegmental vasculature [30]. In cases of incidental SSPE, the data are conflicting, as discussed in Section 4.2.3. Some investigators suggest observation without the use of anticoagulants in this group.

CAT is a common complication of cancer, associated with high morbidity and mortality rates. Once CAT is confirmed, physicians should consider anticoagulant therapy regardless of whether or not a patient has thrombosis-related symptoms. However, limited life expectancy with advanced cancer, high bleeding risks especially in severe thrombocytopenia, or thrombotic symptoms that cannot be expected to improve are factors that prevent the initiation of anticoagulation.

In accordance with major clinical trials comparing LMWH with VKA [14, 31], LMWH was previously the standard treatment for CAT, owing to its efficacy against thrombosis recurrence and improved safety profile compared to VKA. Recently, DOACs have emerged as potential alternative therapies to LMWH owing to their convenient route of administration and predictable pharmacokinetics. However, evidence for their use in CAT is inconclusive, as only a small fraction of the study populations in these trials had CAT [32]. Recently, a Korean phase IV study with rivaroxaban [6] and 2 phase III trials comparing DOACs to LMWH in patients with CAT [13, 33] reported comparable efficacies with slightly increased bleeding risk for DOACs. Based on the results of these 2 randomized controlled trials, the latest guidelines included rivaroxaban and edoxaban as options for VTE treatment [34]. Careful choice of treatment candidates and proper anticoagulant strategies are critical for the treatment of CAT [34, 35]. For anticoagulation in patients with CAT and thrombocytopenia, see the subsection ‘Anticoagulation in thrombocytopenia’.

Treatment of superficial venous thrombophlebitis: Although superficial thrombophlebitis is considered benign and self-limiting, there is increasing recognition that a significant proportion of those presenting with this disorder will have concomitant DVT or PE or are at a significant risk of developing VTE [36]. The treatment of superficial thrombophlebitis remains controversial. However, therapeutic strategies must include symptomatic relief, limitation of thrombosis extension, and reduction in PE risk. A randomized controlled trial showed that anticoagulation with UFH, LMWH, and VKA were superior to compression therapy alone in reducing superficial thrombophlebitis extension [37, 38]. The SURPRISE study suggested the usability of an oral anticoagulant (rivaroxaban) in patients with superficial thrombophlebitis [39]. Current guidelines recommend treating superficial thrombophlebitis of at least 5 cm in length with prophylactic doses of fondaparinux or LMWH for 45 days, but guidelines state that for patients with less extensive superficial thrombophlebitis, anticoagulant treatment is not necessary [16].

Treatment of catheter-related venous thrombosis: Thrombotic complications associated with the use of catheters are common in patients with cancer, leading to patient distress, catheter dysfunction, and infection. There are no large, prospective, randomized trials of treatments for catheter-related thrombosis [40]. Symptomatic catheter-related thrombosis is treated with anticoagulation alone, generally without catheter removal. The optimal duration of anticoagulation for catheter-related thrombosis has not been investigated in clinical trials. The American College of Chest Physicians (ACCP) guidelines recommend anticoagulation for 3 months if the catheter has been removed and for as long as the catheter is in place if >3 months [16]. Recently, DOACs can reportedly be used in cancer patients with catheter-related thrombosis [41, 42]. Multiple reviews of strategies for preventing catheter-related thrombosis are available [43, 44].

Treatment of splanchnic vein thrombosis: Splanchnic vein thrombosis (SVT) is the most common form of unusual site thrombosis, comprising approximately 4% of all thrombotic events. SVT encompasses thrombosis in the portal, mesenteric, hepatic, and splenic veins [45]. Doppler ultrasonography is the first-line choice for the diagnosis of portal and hepatic vein thrombosis, and CT angiography is the diagnostic modality of choice for detecting mesenteric vein thrombosis [46]. The most common acquired risk factors for SVT are abdominal cancers (especially hepatobiliary, gastrointestinal, and pancreatic), liver cirrhosis, and myeloproliferative neoplasms (MPNs). Thus, in non-cirrhotic, non-malignant SVT patients, screening for MPNs (i.e., JAK2 mutation analysis) and paroxysmal nocturnal hemoglobinuria is recommended, even in patients with normal complete blood cell counts [47]. No high-quality randomized controlled trial has yet been reported, but several guidelines or reviews have suggested the management of SVTs [48-51]. Acute, symptomatic, extensive SVTs, and patients planning to undergo liver transplantation are recommended to undergo anticoagulation over surveillance. Esophageal or fundic varices, especially in cirrhotic patients, are not a contraindication for anticoagulation, but adequate prophylactic treatment (with β-blockers, endoscopic ligation, etc.) should precede anticoagulation. To date, there is no data on DOACs for the treatment of SVTs. UFH, LMWH, and heparin transition to VKA are recommended for SVTs. At least 3 months of anticoagulation is recommended for all SVT patients; patients with transient risk factors (surgery, infections) may discontinue treatment after 3 months of anticoagulation, but patients with cirrhosis, active cancer, MPNs, and other thrombophilic conditions should undergo extended anticoagulation while weighing its risks and benefits. In every circumstance, gastrointestinal bleeding should be monitored before or during anticoagulation and should be properly controlled.

Although its absolute risk is low (0.6–1.2 of every 1,000 deliveries) [52], VTE is an important issue in pregnant women, as it is one of the most common causes of maternal morbidity and mortality [53]. Among VTE cases, PE is a particularly important cause of maternal death. For pregnant women with suspected PE, undergoing computed tomographic pulmonary angiography (CTPA) or ventilation-perfusion scan is often inevitable because of the limited diagnostic specificity and sensitivity of D-dimer. A recently published pregnancy-adapted YEARS algorithm for the diagnosis of suspected PE can be used to rule out PE safely; in a prospective study, 32–65% of patients avoided unnecessary CTPA scanning, with the highest efficiency during the 1^st trimester (Fig. 2) [54].

Pregnant women with acute VTE should receive anticoagulation. VKA is contraindicated due to the risk of fetal malformations. The safety of DOACs is not supported by the literature, and concerns have been raised regarding fetal safety in some studies [55]. LMWH is preferred over UFH in terms of convenience.

In times of delivery, a scheduled delivery is suggested if a therapeutic dose of LMWH is administered, based on a 1.9-fold increased risk of bleeding in spontaneous labor compared to planned delivery [56]. In contrast, spontaneous labor is allowed if a woman receives only a prophylactic dose of LMWH [53]. For breastfeeding women, UFH, LMWH, and VKA are considered safe anticoagulants, while DOACs are not yet recommended because of insufficient data [53].

Cause and risk factors of venous thromboembolism in children: Venous thrombosis in children is rare and is mostly associated with multifactorial medical conditions [57]. The incidence of symptomatic VTE in neonates is 0.51 per 10,000 [58], and in children, 0.07 to 0.14 per 10,000 [59, 60]. In the Canadian registry, 96% of DVT/PE cases were associated with medical conditions, including cancer, congenital heart disease, or trauma. The single most important and frequent risk factor for VTE is a central venous catheter (CVC) [61], found in 33–48% of VTE cases in children [59, 61]. The rate of CVC as a risk factor is as high as 94% in neonates [59]. In contrast, inherited thrombophilia was reported in only 8.8% of the cohort [61]. Factor V Leiden and prothrombin gene mutations are frequent in Caucasian patients but not in those of Asian ethnicity. The incidence of protein C deficiency is 0.2–0.5% in the general population but 2–5% in VTE individuals [62-65]. Measurement of protein S is very challenging, and the incidence of deficiency is thought to be as low as 0.9% [66]. The prevalence of hereditary antithrombin III deficiency is estimated to be 0.03–0.8% [67].

The utility of diagnostic tests for inherited thrombophilia in children: A panel of tests for thrombophilia should include factor V Leiden mutations; prothrombin 20210 mutations; deficiencies of antithrombin, protein C, and protein S; homocysteine levels; factor VIII levels; lipoprotein levels; and detection of antiphospholipid antibodies [68]. The developmental stage of the hemostatic system in children, physiologically lower levels of coagulant and anticoagulant proteins, and other confounding factors, such as concomitant use of drug, infection, or inflammation, should be addressed [68-70]. Generally, the tests for thrombophilia do not change decisions about the initiation and duration of acute management, except for specific patients with purpura fulminans complicating severe protein C or S deficiency or cases with VKA-induced skin necrosis [71-73].

Testing in patients with unprovoked, spontaneous, or recurrent thrombotic events, which are rare in children and adolescents, may help to identify inherited thrombophilia and guide management [57, 68]. Up to half of VTE events at unusual sites, such as the cerebral or splanchnic veins, and stroke are associated with inherited thrombophilia, and the patients are candidates for testing [74, 75]. Therefore, testing after the 1^st episode of CVC-related VTE without a strong family history (a first-degree relative with VTE in patients under 40 years old) is not generally recommended [76].

Treatment of venous thromboembolism in children: Treatment recommendations for pediatric VTE are based on studies from adult populations, as scarce evidence exists in children [77]. Treatment options for VTE in children include anticoagulation, thrombolysis, surgery, and observation [78]. Potential benefits must be weighed against risks in premature neonates and critically ill children who are at high risk of bleeding [79]. Options for acute anticoagulation include UFH or LMWH, but LMWH is more frequently used because of the ease of dosing and the low need for monitoring [79]. After an acute period of anticoagulation with UFH or LMWH, patients may continue to receive LMWH, or a transition may be made to VKA [77]. When starting a patient on VKA, UFH or LMWH, admnistration should be continued until the international normalized ratio (INR) reaches the therapeutic range and remains there for 2 days. In most cases, this process takes 5 to 7 days [77]. Anticoagulation with LMWH or VKA, rather than no anticoagulation, is recommended in children with symptomatic DVT or PE [77, 78]. It is suggested to use anticoagulation for ≤3 months in children with provoked DVT or PE and for 6 to 12 months in unprovoked DVT or PE [77]. Patients with recurrent thrombosis and antiphospholipid antibody syndrome often require indefinite anticoagulation [72]. Therapeutic UFH in children is titrated to achieve an activated partial thromboplastin time range (aPTT) of 1.5 to 2.5 times the upper limit of normal [77]. For neonates and children receiving either once- or twice-daily therapeutic LMWH, the drug needs to be monitored with a target anti-Xa activity range of 0.5 to 1.0 unit/mL in a sample taken 4 to 6 h after subcutaneous injection [77]. For children receiving VKA, the drug needs to be monitored to a target INR of 2.5 (range, 2.0–3.0), except for those with prosthetic cardiac valves, who are recommended to adhere to the adult guidelines [77]. Symptomatic CVC-related thrombi can be treated without removal if the catheter is functioning. A nonfunctioning or unnecessary catheter should be removed after 3 to 5 days of therapeutic anticoagulation. The total recommended anticoagulant duration is between 6 weeks and 3 months [77]. Thrombolytic therapy should be used only for life- or limb-threatening thrombosis [78]. Thrombectomy followed by anticoagulation or IVC filter insertion is not recommended except in specific, limited circumstances; rather, anticoagulation alone should be used in children with symptomatic DVT or PE [78]. DOACs have potential benefits for children, including oral administration, and possibly no need for monitoring. Many clinical trials of DOACs are underway, but the routine use of DOACs is not yet recommended in children [80].