Prevention of venous thromboembolic disease in adult nonorthopedic surgical patients

INTRODUCTION — Venous thromboembolism (VTE; deep venous thrombosis and pulmonary embolism [PE]) is common in the postoperative setting with over half of this population at moderate risk for VTE [1-3]. PE is one of the most common preventable causes of in-hospital deaths following surgery [4-8].

Nonorthopedic surgeries include surgery of the skin and soft tissues of the trunk or extremities; surgery involving the chest, abdomen, or pelvic organs; and surgery of the head (including brain) and neck. Our approach to the prevention of VTE in nonorthopedic surgical patients will be reviewed here. Our approach is, for the most part, in keeping with recommendations from several societies, including the American College of Chest Physicians (ACCP) and the American Society of Hematology (ASH) [9-11]. Prevention of VTE in patients undergoing orthopedic procedures (eg, joint repair/replacement) and in hospitalized medical patients are presented separately. (See "Prevention of venous thromboembolism in adults undergoing hip fracture repair or hip or knee replacement" and "Prevention of venous thromboembolic disease in acutely ill hospitalized medical adults".)

ASSESS RISK FOR THROMBOSIS — The risk of postoperative VTE should be assessed prior to surgery and the patient stratified into very low, low, moderate, or high-risk groups so that an appropriate method of VTE prevention can be selected [12]. (See 'Baseline thrombosis risk' below and 'Thrombosis risk model (Caprini)' below.)

VTE risk largely depends upon the procedure but patient-related factors also play a role:

●Procedure-related – Many procedure-related factors contribute to the risk of VTE in nonorthopedic patients including the extent and duration of surgery, intraoperative positioning, the type of anesthesia, and postoperative mobility. In general, the highest risk is in those undergoing major surgery (defined as surgery lasting longer than 45 minutes [13]), abdominal and thoracic cavity surgery (eg, major abdominal/pelvic or surgery for malignancy), prolonged surgery (≥2 hours), emergency rather than elective surgery, postoperative immobilization for ≥4 days, as well as critically ill patients who are confined to bed (eg, extensive burns, multiple trauma, brain/spine injury) [13-16]. The risk is generally low for patients undergoing minor, typically ambulatory procedures (eg, elective hernia repair, thyroid surgery, minor skin excision, carotid endarterectomy).

●Patient-related – Patient-related VTE risk factors are discussed separately. (See "Overview of the causes of venous thrombosis".)

Baseline thrombosis risk — The baseline risk of VTE associated with individual surgeries is highly variable reflecting the wide range of surgeries within each specialty [12,17]. Importantly, estimates of baseline risk are imperfect since they use data from studies in populations not on VTE prophylaxis and populations on prophylaxis, or use extrapolated data from related populations.

General, abdominal/pelvic, bariatric, vascular, plastic surgery — The risk categories (very low, low, moderate, high) correspond to the Caprini model (table 1) discussed below, although in practice many experts use clinical gestalt to estimate the risk. (See 'Thrombosis risk model (Caprini)' below.)

●General/abdominal pelvic (low to high Caprini score) – Rates of symptomatic VTE derived from untreated groups in randomized trials [18-30] have ranged from 0.5 to 1.6 percent, higher among those undergoing surgery for malignancy (up to 3.7 percent) [6,18]. However, several studies report that VTE risk is wide in this population since it encompasses a broad range of surgeries from laparoscopic appendectomy to open pelvic surgery for cancer.

●Bariatric surgery (low to high Caprini score) – Observational data suggest that while in the past rates were between 1.9 and 5.4 percent [31,32], advances in less extensive bariatric surgery may be associated with lower rates (0.5 percent) according to data from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) [33].

●Noncardiac vascular surgery (low to high Caprini score) – Baseline risk in the absence of prophylaxis is extrapolated from the population of patients undergoing abdominal/pelvic operations since vascular procedures were commonly included in those studies. Small observational studies comparing pharmacologic prophylaxis with no prophylaxis in only vascular surgery patients report variable rates for open abdominal vascular surgery (up to 10 percent) [34-41], peripheral artery surgery (1.8 to 9 percent) [42-46], venous ablation procedures (<1 percent) [47-57], and lower extremity amputation (2 to 15 percent, higher for above knee compared with below-knee) [37,58-62].

●Plastic and reconstructive (low to high Caprini score) – In the absence of pharmacologic prophylaxis, the baseline risk as estimated by the American College of Chest Physicians (ACCP) is between 0.5 to 1.8 percent based upon three observational studies [12,63-65]. However, extrapolating from related surgical populations (breast surgery, lower extremity bypass surgery), the estimated baseline VTE risk ranges from low (eg, outpatient cosmetic procedures) to high (eg, reconstructive surgeries).

Cardiac, thoracic, neurosurgery, major trauma — The ACCP has provided rough estimates of the baseline risk for VTE, in the absence of prophylaxis, for the other nonorthopedic surgical specialties [12]. The risk categories in parenthesis below correspond to the Caprini model, although this scoring system (table 1) has not been validated in these populations, and most experts use gestalt assessment. (See 'Thrombosis risk model (Caprini)' below.)

●Cardiac surgery (moderate to high Caprini score) – Several studies have identified rates of VTE up to 1 percent in this population (prophylaxis unknown) [6,66,67] but older studies suggest higher rates (up to 25 percent) in the absence of prophylaxis [68-70].

●Noncardiac thoracic surgery (moderate to high Caprini score) – Several studies reported that the incidence of symptomatic VTE ranges from 0.18 to 7.4 percent (highest in pneumonectomy, esophagectomy, extended resection) [6,66,71-73].

●Neurosurgery (moderate to high Caprini score) – Meta-analyses report a pooled incidence of VTE in untreated patients between 16 and 29 percent, highest in those undergoing craniotomy [29,30]. In studies limited to spinal surgery, the incidence of deep vein thrombosis (DVT) ranges from 0 to 15 percent (with and without prophylaxis) [74-80]. Limited spine surgeries for benign conditions and cervical spine surgery are associated with less risk [74,75].

●Major trauma (moderate to high Caprini score) – While studies report an incidence of DVT as high as 58 percent among those not receiving prophylaxis [81], these rates may reflect the most seriously ill patients with multiple other injuries (eg, traumatic brain and spinal injury) [82-87]. However, a 2013 systematic review, reported lower incidences of DVT and PE in patients who received either no prophylaxis (8.7 percent) or only mechanical prophylaxis (3.7 percent) [88]. Accurate estimates have been hampered by a baseline risk that varies widely in this population since these patients often undergo abdominal, vascular, neurologic, and/or orthopedic surgeries [81-87,89-91].

Thrombosis risk model (Caprini) — Although there have been many attempts to quantitate VTE risks, no one method has been found to be universally acceptable and many physicians use a gestalt assessment [12-14,92-94]. Nonetheless, the most widely used model is the Modified Caprini Risk Assessment Model (ie, Caprini score modified by the ACCP (table 1) [12]). The Rogers score is less frequently used and has not been externally validated [14].

Using the Caprini score, patients undergoing surgical procedures are classified according to their estimated baseline risk (EBR) for VTE in the absence of thromboprophylaxis as (see 'Baseline thrombosis risk' above):

●Very low risk – Caprini score 0; corresponding to an EBR <0.5 percent (see 'Very low thrombosis risk: Early ambulation' below)

●Low risk – Caprini score 1 to 2; corresponding to an EBR of about 1.5 percent (see 'Low VTE risk: Mechanical methods' below)

●Moderate risk – Caprini score 3 to 4; corresponding to an EBR of about 3 percent (see 'Moderate or high VTE risk' below)

●High risk – Caprini score ≥5; corresponding to an EBR of at least 6 percent (see 'Moderate or high VTE risk' below)

The caveat of this model is that it has been validated, and is therefore, only applicable to patients undergoing general (eg, breast, thyroid, parathyroid) and abdominal/pelvic surgery (eg, gastrointestinal, urologic, gynecologic), including those who are critically ill [12,92,94-96]. Although not validated in other populations, it is considered by most experts as acceptable for use in those undergoing bariatric and vascular surgery. In addition, this model underwent further modification for patients undergoing plastic/reconstructive surgery since a validation study reported a lower risk of VTE for a given Caprini score in this population (0.6 percent among those with a score of 3 to 4, 1.3 percent with a score of 5 to 6, 2.7 percent with a score 7 to 8, and 11.3 percent with a score of >8).

ASSESS RISK FOR MAJOR BLEEDING — For patients in whom pharmacologic VTE prophylaxis is indicated, a full history and examination should be obtained to assess the risk for major bleeding. Major bleeding is defined as fatal bleeding, and/or symptomatic bleeding in a critical area or organ (perhaps requiring reexploration), and/or bleeding causing a fall in hemoglobin of ≥2 g/dL or leading to transfusion of two or more units of whole blood or red cells [97]. When assessing the risk, we prefer that the baseline risk be assessed first and then modified according to the potential consequences of bleeding and individual risk factors discussed below. (See 'Estimates of baseline bleeding risk' below and 'Individual risk factors for bleeding' below and 'Bleeding risk categories' below.)

The rate of bleeding associated with pharmacologic prophylaxis varies among patient groups. One meta-analysis of 51 randomized trials of pharmacologic VTE prophylaxis in general surgery patients reported that minor bleeding was common and included injection site bruising (7 percent), wound hematoma (6 percent), drain site bleeding (2 percent), and hematuria (2 percent) [98]. Major bleeding complications were uncommon and included including gastrointestinal tract (0.2 percent) or retroperitoneal (<0.1 percent) bleeding. Discontinuation of prophylaxis occurred in 2 percent of patients and subsequent reoperation for bleeding occurred in less than 1 percent.

Estimates of baseline bleeding risk — Baseline bleeding risk has been poorly studied in nonorthopedic surgical patients. Although several studies have attempted to elucidate bleeding risk [12,99], risk stratification for major bleeding has been estimated by the American College of Chest Physicians (ACCP) in the following patient groups as [12]:

●General/abdominal/pelvic surgery – 1 percent

●Bariatric surgery – <1 percent

●Plastic and reconstructive surgery – 0.5 to 1.8 percent

●Vascular surgery – 0.3 to 1.8 percent

●Cardiac surgery – 5 percent (high risk)

●Thoracic surgery – 1 percent

●Neurosurgery – Craniotomy: 1 to 1.5 percent; spinal surgery: <0.5 percent

●Major trauma – 3.4 to 4.7 percent (high risk)

Individual risk factors for bleeding — Patients with individual risk factors for bleeding include those with active bleeding as an indication for surgery (eg, gastrointestinal bleeding, trauma, ruptured aneurysm), patients with intracranial hemorrhage, patients who develop a moderate or severe coagulopathy (eg, patients with liver disease), and patients with an underlying bleeding disorder or thrombocytopenia (eg, platelet count <50,000/microL, or <100,000/microL plus additional risk factors for bleeding). Patients with relative contraindications include those with recurrent bleeding from multiple gastrointestinal telangiectasias. Epistaxis and menstrual bleeding are not contraindications to pharmacologic thromboprophylaxis. (See "Prevention of venous thromboembolic disease in acutely ill hospitalized medical adults", section on 'Bleeding risk assessment'.)

Bleeding risk categories — Following the above assessment, bleeding risk can be categorized as either low or high.

●Low bleeding risk – In general, patients undergoing general, abdominal-pelvic, bariatric, vascular, and thoracic surgery that is uncomplicated tend to have lower rates of bleeding (<2 percent) when compared with other patients.

●High bleeding risk – Patients undergoing cardiac surgery and patients with major trauma, especially involving the brain and spine [80], are at highest risk of bleeding (>3 percent). Patients in this category also include those in whom the consequences of bleeding are considered potentially devastating; for example, patients undergoing neurosurgical procedures where thromboprophylaxis may result in spinal or intracranial hemorrhage and patients undergoing plastic/reconstructive surgery where thromboprophylaxis may result in injury or rejection of grafted tissue due to bleeding. Similarly, patients with one or more individual risk factors for bleeding are considered at high risk of bleeding postoperatively. Prophylaxis in this population is discussed below. (See 'With high bleeding risk: Mechanical methods' below.)

SELECTING THROMBOPROPHYLAXIS — Options for primary VTE prophylaxis include early ambulation, pharmacologic and/or mechanical methods. VTE prevention strategies should be individualized according to the risk of VTE (very low, low, moderate, and high) as well as the risk and consequences of major bleeding. (See 'Assess risk for thrombosis' above and 'Assess risk for major bleeding' above.)

The approach outlined in the sections below is, in general, consistent with international guidelines including the American Society of Hematology (ASH), American College of Chest Physicians (ACCP), the Asian Venous Thrombosis Forum, Korean guidelines for the Prevention of Venous Thromboembolism, European guidelines on perioperative venous thromboembolism prophylaxis: Executive summary, and the International Consensus Statement on the Prevention and Treatment of Venous Thromboembolism [12,100-106]. Importantly, the approach assumes that patients are at low risk for bleeding. In addition, the decision is fluid such that individual circumstances before, during, and after surgery may alter the decision regarding method selection. In addition, individualizing the approach according to individual factors is also prudent. Clinicians should also be aware that in general, thromboprophylaxis reduces but does not eliminate VTE events and VTE-related mortality [26,66,107,108].

Noteworthy, methods of secondary prophylaxis (eg, surveillance imaging) and inferior vena cava filters are not recommended for VTE prevention in this population.

Various strategies to improve the use of thromboprophylaxis methods have been demonstrated to be effective, including computerized order sets with electronic alerts, or preprinted orders and quality improvement in the form of clinician education programs, audit, and feedback [3,109-115]. Further efforts are required to improve the use of thromboprophylaxis in clinical practice.

Very low thrombosis risk: Early ambulation — The risk of VTE is considered very low when the baseline risk in the absence of prophylaxis is estimated to be less than 0.5 percent (table 1). Very low-risk group surgeries generally include patients undergoing general or abdominal/pelvic surgery with a Caprini score of zero or patients undergoing plastic/reconstructive surgeries with a Caprini score of zero to two. Examples include healthy young patients undergoing minor outpatient procedure (eg, LASIK surgery, cataract removal, skin biopsy, benign breast biopsy, diagnostic endoscopy, nasal polyp removal, dilatation and curettage, colposcopy, fluid removal from joint effusion).

For nonorthopedic surgical patients at very low risk of VTE, we recommend early and frequent ambulation rather than pharmacologic or mechanical methods of prophylaxis. Most very low-risk patients are able to ambulate easily after surgery. Mechanical methods may be employed in the unusual circumstance where unexpected issues occur during the procedure (eg, bleeding, more extensive surgery, which intrinsically change the risk category) or the patient has a complication and requires admission.

Since there are no randomized trials comparing ambulation with other methods, this approach is largely based upon the rationale that the baseline rate of VTE in this population is too low (<0.5 percent) to warrant prophylaxis. In addition, indirect data from studies evaluating the risk of VTE in orthopedic patients have also suggested that the risk of VTE is lowered by 70 percent in those who ambulate on or before the second postoperative day [116]. (See "Prevention of venous thromboembolism in adults undergoing hip fracture repair or hip or knee replacement", section on 'Risk of thrombosis'.)

Low VTE risk: Mechanical methods — The risk of VTE is considered low when the baseline risk in the absence of prophylaxis is estimated to be 1.5 percent (table 1). Patients in this category include those undergoing general or abdominal/pelvic surgery with a Caprini score of 1 to 2 or patients undergoing plastic/reconstructive surgery with a Caprini score of 3 to 4. Examples include those undergoing minor elective abdominal-pelvic surgery (eg, appendectomy, laparoscopic cholecystectomy) or minor thoracic surgery (eg, diagnostic thoracoscopy, video-assisted biopsy). Other examples include patients undergoing minor vascular procedures (eg, vein ablation), and elective spine surgery (eg, spinal fusion [117]). (See 'Baseline thrombosis risk' above.)

For nonorthopedic surgical patients at low risk for VTE, we suggest mechanical methods of VTE prophylaxis rather than pharmacologic prophylaxis or no prophylaxis. The rationale for this approach is that the risk of VTE is high enough to justify thromboprophylaxis but does not warrant the risk of bleeding associated with pharmacologic methods. Switching to pharmacologic methods may be appropriate in this with individual risk factors for VTE (eg, history of recurrent VTE or cancer).

Mechanical methods of thromboprophylaxis include intermittent pneumatic compression (IPC), graduated compression stockings (GCS, also known as elastic stockings), and the venous foot pump (VFP). Although IPC devices may be superior and are preferred by the ACCP, experts generally choose IPC or GCS since the data comparing one device over the other is fundamentally flawed [12,118-120].

Intermittent pneumatic compression and venous foot pump — Data supporting the use of IPC for the prevention of VTE in nonorthopedic surgical patients are limited. However, of the mechanical devices (IPC, GCS, VFP), efficacy appears best with IPC use (picture 1) [118,120-122]. Meta-analyses of small randomized trials of mixed surgical populations (including general, abdominal, urologic, neurosurgery, oncologic, orthopedic) report that IPC use is superior to no prophylaxis and to GCS, and may offer additive benefit to surgical patients on low molecular weight (LMW) heparin [12,121,123-127]. (See 'With low bleeding risk: Combined prophylaxis' below.)

As an example, the largest meta-analysis, which included data on 16,164 patients (mostly surgical) enrolled in 70 trials, reported that IPC was more effective than no prophylaxis in reducing deep venous thrombosis (DVT; 7.3 versus 16.7 percent) and pulmonary embolism (PE; 1.2 versus 2.8 percent) without any effect on mortality [123]. Although the addition of pharmacologic prophylaxis to IPC further reduced the risk of DVT (relative risk [RR] 0.54, 95% CI 0.32-0.91), it had no effect on the incidence of PE. In another meta-analysis of eight studies that compared mechanical prophylaxis with LMW heparin, the risk of DVT (symptomatic and asymptomatic) was 80 percent higher in those who had mechanical methods of prophylaxis (RR 1.80, 95% CI 1.16-2.79) [122] with a 57 percent decrease in the risk of major bleeding. However, patients in these studies had a range of VTE risk and many were of moderate to high risk.

Randomized studies showing efficacy of VFP devices (picture 2) in surgical patients are lacking but, similar to IPC devices, they prevent thrombosis by stimulating lower limb venous flow [128,129].

Compliance, proper fit, and discomfort are major issues with IPC devices. Devices may be removed while the patient is ambulating but should be put back on when the patient returns to a seated or supine position. However, reflective of practice, observational studies report frequent errors in IPC application [130-132]. This suggests that frequent interference with the device is common and may potentially interfere with efficacy. Battery-operated devices may improve compliance in the future.

IPC is contraindicated in patients with evidence of leg ischemia (eg, peripheral artery disease). Although there are no data available on skin complications of IPC use, skin breakdown is a known complication, especially in frail, older adults, although some clinicians use loose stockinettes underneath the device to counteract this phenomenon. In addition, practical considerations for amputees or patients with burns or extensive skin lesions (eg, Stevens Johnson's syndrome) may limit IPC application. In this context, although one device can be applied to any extremity, its efficacy in the prevention of VTE is not assured and likely limited. There is also a hypothetical concern that patients who have been immobilized for a period of ≥72 hours without any form of prophylaxis may be at risk of dislodging recently formed venous clot in the lower extremities. The value of clinical examination or ultrasound in determining risk of clot dislodgement following the application of IPC in this setting is unknown.

Optimal timing of IPC in surgical patients is poorly studied. However, one study suggested that IPC should be started as soon as possible, preferably just before surgery or in the operating room and continued with few interruptions until discharge [133]. (See 'Timing of initiation' below.)

IPC devices are thought to prevent VTE by enhancing blood flow in the deep veins of the legs, thereby preventing venous stasis [134]. IPC also reduces plasminogen activator inhibitor-1 (PAI-1), thereby increasing endogenous fibrinolytic activity [135]. (See "Thrombotic and hemorrhagic disorders due to abnormal fibrinolysis", section on 'PAI-1'.)

Graduated compression stockings — There is a paucity of high quality randomized trials that have studied the efficacy of GCS for preventing VTE in the surgical population [120,136-141]. GCS alone are effective at preventing DVT but may be less effective than pharmacologic agents. However, GCS when combined with other prophylactic methods appears to improve rates of DVT prevention. Of note, GCS refer to prescription-style stockings and not anti- embolism stockings, which are widely available in most facilities. (See 'With low bleeding risk: Combined prophylaxis' below.)

As examples:

●One meta-analysis of 20 randomized trials mostly surgical and orthopedic patients (mixed populations) reported that the use of GCS alone was more effective than no prophylaxis in the prevention of DVT (21 versus 9 percent; odds ratio [OR] 0.35, 95% CI 0.28-0.43) [142].

●In another meta-analysis of eight studies (mostly hospitalized general surgical and orthopedic patients), the incidence of proximal DVT and PE was lower in patients treated with GCS compared with patients without GCS (1 versus 5 percent, 2 versus 5 percent, respectively) [141]. However, addition of a second method of prophylaxis in some of the included trials may have biased the favorable outcome associated with GCS in this analysis.

●Trials reporting the efficacy of GCS when combined with pharmacologic prophylaxis are discussed below. (See 'With low bleeding risk: Combined prophylaxis' below.)

The efficacy and safety of thigh- versus knee-length stockings has not been studied in surgical patients.

Optimal timing of GCS in surgical patients is unstudied. However, in general, GCS should be started as soon as possible, preferably before surgery, or in the operating room and continued with few interruptions until discharge. (See 'Timing of initiation' below.)

Contraindications against their use and local skin breakdown complications are similar to those for IPC. (See 'Intermittent pneumatic compression and venous foot pump' above.)

Moderate or high VTE risk — The risk of VTE is considered moderate when the baseline risk in the absence of prophylaxis is estimated to 3 percent and high if it is at least 6 percent (table 1). (See 'Baseline thrombosis risk' above.)

●Moderate-risk surgical patients have been defined as patients undergoing general or abdomen/pelvic surgery with a Caprini score of 3 to 4 or patients undergoing plastic/reconstructive surgery with a Caprini score of 5 to 6 (table 1). Patients undergoing major gynecologic and urologic surgery usually fall into the moderate risk category [143]. In addition, patients undergoing major cardiac or thoracic surgery, bariatric surgery, and neurosurgical procedures, and patients with nonextensive trauma not involving the brain or spine are, at minimum, also considered moderate risk for VTE.

●High-risk surgical patients have been defined as patients undergoing general or abdominal/pelvic surgery with a Caprini score of 5 or more (table 1) or patients undergoing plastic/reconstructive surgery with a Caprini score of 7 to 8. Examples of patients in the high-risk group are those undergoing extensive thoracic or abdominal-pelvic surgery (eg, distal colorectal surgery, extensive pelvic surgery, lung cancer resection, esophagectomy, brain cancer resection), major trauma (particularly if involving the brain or spinal cord), acute spinal cord injury, or cancer surgery [144,145].

With low bleeding risk: Pharmacologic alone — In nonorthopedic surgical patients at moderate risk for VTE and in whom the risk of bleeding is low, we suggest pharmacologic prophylaxis, rather than mechanical methods, while in those at high VTE risk we recommend pharmacologic prophylaxis rather than mechanical methods. The rationale for this approach is based upon randomized trials and meta-analyses with data for patients at moderate risk being weaker than for those at high risk. For select patients in whom the risk of VTE is considered to be particularly high, we suggest the addition of mechanical to pharmacologic methods (eg, multiple risk factors, surgery for cancer). (See 'With low bleeding risk: Combined prophylaxis' below and 'Pharmacologic dosing' below.)

Among the available agents, low molecular weight (LMW) heparin is generally the preferred anticoagulant based upon randomized trials that report superior or similar efficacy with unfractionated heparin (UFH) or fondaparinux, although most data show no appreciable effect on mortality and limited effect on clinically relevant bleeding. For those with renal insufficiency (creatinine clearance <20 to 30 mL/min) or for those in whom cost is an issue, UFH is appropriate (table 2). For patients in whom UFH or LMW heparin is contraindicated (eg, heparin-induced thrombocytopenia [HIT]) or unavailable, fondaparinux or mechanical methods are preferred. Timing of initiation and dosing of these agents are discussed below. (See 'Administration' below.)

Many experts administer more aggressive prophylaxis in very high-risk populations in the form of increased intensity of a pharmacologic agent (eg, three times a day UFH, twice daily enoxaparin) and/or the addition of a mechanical device (usually IPC). Patients in this category include those undergoing abdominal pelvic surgery with a Caprini score >8, patients with multiple risk factors, patients undergoing craniotomy or spinal surgery for cancer, and patients with major trauma, especially that involving the brain or spine. (See 'With low bleeding risk: Combined prophylaxis' below.)

The efficacy of pharmacologic therapy in comparison with mechanical methods, and specifically for LMW heparin compared with other anticoagulants is summarized below:

●Pharmacologic versus no prophylaxis or mechanical methods – In general, randomized trials and meta-analyses have shown that pharmacologic prophylaxis with LMW heparin, low-dose UFH, and fondaparinux are superior to placebo or mechanical devices. However, in general, the level of support from data in patients at moderate risk is weaker than for those at high risk.

•LMW heparin – In a meta-analysis of eight trials of five different preparations of LMW heparin that included >48,000 general and abdominal surgery patients, compared with no prophylaxis, LMW heparin reduced the risk of symptomatic VTE by 70 percent but resulted in a doubling of the risk of major bleeding and wound hematomas [18].

Similar results were obtained from another meta-analysis that also included gastrointestinal, urologic, gynecologic, and thoracic surgery patients [146].

Trials reporting the efficacy of LMW heparin combined with compression stockings in nonorthopedic patients including those undergoing neurosurgery are discussed below. (See 'With low bleeding risk: Combined prophylaxis' below.)

•Low-dose UFH – An early trial involving over 4000 patients established the efficacy of low-dose UFH for reducing the incidence of fatal PE in patients undergoing major surgical procedures compared with controls (0.7 versus 0.1 percent) [107]. Pooled data from meta-analyses subsequently confirmed that low-dose UFH reduced the incidence of all DVT, proximal DVT, and all PE including fatal PE, when compared with placebo [26,98,147,148].

●LMW heparin versus other agents – Our preference for LMW heparin is based upon direct data derived from nonorthopedic surgical populations that included randomized trials and meta-analyses, most of which show similar or superior efficacy compared with UFH as well as indirect data that show similar benefits in orthopedic and medical populations. However, the quality of data is limited by the heterogeneity of study populations eg, (wide range in risk, different surgery types) and the failure to distinguish symptomatic from asymptomatic VTE. (See "Prevention of venous thromboembolism in adults undergoing hip fracture repair or hip or knee replacement" and "Prevention of venous thromboembolic disease in acutely ill hospitalized medical adults", section on 'Low molecular weight heparin'.)

●LMW heparin versus low-dose UFH – Low-dose UFH is generally considered an alternative to LMW heparin, when, for example, cost or renal insufficiency is an issue. While early meta-analyses comparing low-dose UFH and LMW heparin found similar efficacy and safety for both agents [149,150], newer analyses suggest that LMW heparin is superior. As an example, compared with low-dose UFH, one meta-analysis of 51 randomized trials of general surgery patients reported that the rate of VTE events was 30 percent lower in patients receiving LMW heparin without any effect on death or bleeding [18]. However, the same efficacy was not apparent when the analysis was confined to placebo-controlled trials. Most of the patients in these trials underwent abdominal (particularly for gastrointestinal disease) or thoracic surgery, but some patients underwent gynecologic or urologic surgery, mastectomy, or vascular procedures [26,147].

Similar results of superior or similar efficacy were reported in other meta-analyses and randomized trials that included patients undergoing gynecologic and urologic surgery [18,146,149-152], and cancer surgery [153-159].

For patients admitted with major trauma, LMW heparin was also reported to be superior to UFH in the prevention of both total and proximal DVTs [27,160-162].

●Fondaparinux versus UFH or LMW heparin – Fondaparinux is an alternative to LMW heparin and UFH in patients with contraindications to heparin (eg, HIT) or when these agents are not available. In a randomized trial (PEGASUS) of 2408 patients undergoing major abdominal surgery, fondaparinux and the LMW heparin, dalteparin, had a similar efficacy in reducing the rate of VTE (4.4 versus 6.1 percent) without any increased rate of major bleeding (3.4 versus 2.4 percent) [152]. In a meta-analysis that pooled the results from this study with studies including orthopedic patients, when compared with LMW heparin, fondaparinux did not reduce VTE events but resulted in an increased risk of bleeding [9]. (See "Prevention of venous thromboembolism in adults undergoing hip fracture repair or hip or knee replacement".)

●Oral agents (warfarin, aspirin, direct oral anticoagulants) – Although oral agents including warfarin, aspirin, direct thrombin inhibitors (eg, dabigatran) and factor Xa inhibitors (eg, rivaroxaban, edoxaban, and apixaban) are sometimes administered in orthopedic patients for VTE prevention, they are unstudied and are not routinely administered in nonorthopedic surgical patients. Occasionally, for patients undergoing general or abdominal/pelvic surgery considered at high risk for VTE in whom LMW heparin or UFH is contraindicated or unavailable, some experts administer aspirin as an alternative to fondaparinux or mechanical methods, although the data to support this strategy is indirect and derived from patients undergoing major orthopedic surgery [12]. (See "Prevention of venous thromboembolism in adults undergoing hip fracture repair or hip or knee replacement".)

With low bleeding risk: Combined prophylaxis — Combining pharmacologic and mechanical methods of prophylaxis are generally reserved for those considered to be at the highest risk of VTE. Although there are a paucity of high quality studies comparing combined methods with pharmacologic prophylaxis alone, data in general suggest that combining methods of thromboprophylaxis offers additional protective benefit against the development of VTE [136]. However, the contributions of each method is unclear and most studies are limited by incomplete blinding, small sample size, uncertain concealment, and measurement of surrogate outcomes. No convincing data report further benefits in efficacy when two mechanical methods are combined or added to pharmacologic agents.

As examples:

●In a large meta-analysis of 34 trials, which included data on 14,931 patients (mixed surgical, trauma, orthopedic), the addition of IPC to pharmacologic prophylaxis reduced the incidence of PE from 1.8 to 0.9 percent (OR 0.46, 95% CI 0.30-0.71, low-certainty evidence) and the incidence of DVT from 9.2 to 5.5 percent (OR 0.38, 95% CI 0.21-0.70, high-certainty evidence) [163]. The results did not appear to be biased by the inclusion of orthopedic patients in the analysis, since combined prophylaxis resulted in a similar reduction in PE and DVT in both orthopedic and nonorthopedic surgical patients.

●In a meta-analysis of 25 randomized trials in mixed surgical populations, adding pharmacologic agents to GCS almost halved the rate of postoperative DVT (RR 0.56) while the risk of bleeding almost doubled (RR 1.74) compared with GCS alone [140]. In another meta-analysis, rates of DVT were lower in those treated with GCS plus any other prophylactic method compared with GCS alone (4 versus 13 percent) [139].

●Several randomized trials of patients undergoing neurosurgical procedures reported LMW heparin, when added to GCS, prevented more venographic DVT than GCS alone (17 to 19 percent versus 26 to 32 percent) without an increased risk of bleeding [164,165].

●In a randomized trial (APOLLO) of patients undergoing major abdominal surgery (gastrointestinal, urologic), fondaparinux combined with IPC significantly reduced the rate of VTE when compared with IPC alone (1.7 versus 5.3 percent) [166]. Major bleeding was more frequent after fondaparinux (1.6 versus 0.2 percent), although none of the bleeding events were fatal or involved a critical organ.

●In a "quasi"-randomized trial (IPC SUPER) of patients undergoing major surgery and a Caprini score of ≥9, standard care (low molecular weight heparin plus antiembolism stockings) combined with IPC reduced the rate of asymptomatic DVT (0.5 versus 16.7 percent) compared with standard care alone [167]. However, limitations including lack of blinding and the inclusion of asymptomatic VTE as an endpoint prohibit firm conclusions.

With high bleeding risk: Mechanical methods — For patients with contraindications to pharmacologic prophylaxis (eg, active bleeding, intracranial hemorrhage, bleeding diathesis (table 3)), patients at high risk of bleeding, or patients in whom the consequences of bleeding are thought to be potentially catastrophic (eg, neurosurgical procedures), we suggest mechanical methods rather than no thromboprophylaxis. Among the options, IPC is the typical method used but data to support their use over graduated compression stockings or VFP devices are limited. Vena cava filters should not be routinely used for VTE prevention. (See 'Intermittent pneumatic compression and venous foot pump' above and 'Methods not recommended' below.)

Switching to or adding a pharmacologic agent, such as LMW heparin, should be done as soon as the bleeding risk becomes acceptably low (eg, 48 to 72 hours following neurosurgery) or the bleeding diathesis has been reversed. (See 'Assess risk for major bleeding' above.)

SPECIFIC SURGICAL POPULATIONS — Additional data that support VTE prophylaxis in special surgical populations are discussed separately in the followings topics:

General and abdominal/pelvic surgery

●Bariatric surgery (table 4) (see "Bariatric surgery: Postoperative and long-term management", section on 'Venous thromboembolism')

●Colon surgery (see "Overview of colon resection", section on 'Thromboprophylaxis')

●Abdominal aortic aneurysm repair (see "Open surgical repair of abdominal aortic aneurysm", section on 'Thromboprophylaxis' and "Endovascular repair of abdominal aortic aneurysm", section on 'Thromboprophylaxis')

●Extremity amputation (see "Lower extremity amputation", section on 'Thromboprophylaxis' and "Upper extremity amputation", section on 'Evaluation and preparation')

●Gynecologic surgery (see "Overview of preoperative evaluation and preparation for gynecologic surgery", section on 'Thromboprophylaxis')

Neurosurgery including neurotrauma

●Traumatic brain injury (see "Management of acute moderate and severe traumatic brain injury", section on 'Venous thromboembolism prophylaxis')

●Spinal cord injury (see "Respiratory complications in the adult patient with chronic spinal cord injury", section on 'Venous thromboembolism')

●Surgery for brain tumors (see "Treatment and prevention of venous thromboembolism in patients with brain tumors", section on 'Primary prevention (VTE prophylaxis)')

Anesthesia and perioperative care

●Patients undergoing neuraxial anesthesia or analgesia (see "Neuraxial anesthesia/analgesia techniques in the patient receiving anticoagulant or antiplatelet medication" and "Perioperative management of patients receiving anticoagulants", section on 'Neuraxial anesthesia')

●Critically ill patients including postoperative surgical patients (see "Prevention of venous thromboembolic disease in acutely ill hospitalized medical adults", section on 'Selection of method of prophylaxis')

●Trauma patients (see "Venous thromboembolism risk and prevention in the severely injured trauma patient", section on 'Thromboprophylaxis')

●Patients with stroke (see "Prevention and treatment of venous thromboembolism in patients with acute stroke")

●Patients with cancer (table 5) (see "Risk and prevention of venous thromboembolism in adults with cancer")

●Patients who are pregnant (see "Venous thromboembolism in pregnancy: Prevention")

ADMINISTRATION

Timing of initiation — The optimal timing for initiation of mechanical and/or pharmacologic thromboprophylaxis in nonorthopedic patients is unknown and should be individualized according to factors including timing of surgery (elective or emergency), type and duration of surgery, the estimated risk of bleeding, and baseline risk of VTE. (See 'Assess risk for thrombosis' above and 'Assess risk for major bleeding' above.)

Studies that specifically examine timing of pharmacologic dosing are limited. However, the approach below is in accordance with studies that proved efficacy of pharmacologic agents (see 'Pharmacologic dosing' below), as well as data derived from observational studies in trauma patients (ie, patients with the highest risk of VTE and bleeding). (See "Venous thromboembolism risk and prevention in the severely injured trauma patient", section on 'Specific trauma populations'.)

For most patients in whom thromboprophylaxis is indicated and the risk of bleeding is low, experts agree that mechanical methods may commence just before surgery and that pharmacologic agents should ideally commence within 2 to 12 hours preoperatively. The exception to this rule is fondaparinux, which is typically started six to eight hours after skin closure. In patients not considered suitable candidates for preoperative pharmacologic thromboprophylaxis due to a contraindication to anticoagulants, a high risk of bleeding, or potential catastrophic effect of bleeding, mechanical methods should be employed (typically just before surgery) and pharmacologic agents started or added postoperatively, as soon as hemostasis is achieved and it is considered safe (eg, 2 to 72 hours).

All anticoagulants have a boxed warning regarding the risk of spinal or epidural hematoma in patients receiving neuraxial anesthesia or undergoing spinal puncture. The risk is increased in those with indwelling epidural catheters, other drugs that impair hemostasis (eg, anti-platelet agents), traumatic or repeated epidural or spinal puncture, or a history of spinal surgery. Evidence-based guidelines from the American Society of Regional Anesthesia (ASRA) are discussed separately. (See "Neuraxial anesthesia/analgesia techniques in the patient receiving anticoagulant or antiplatelet medication" and "Perioperative management of patients receiving anticoagulants", section on 'Neuraxial anesthesia' and "Overview of neuraxial anesthesia", section on 'Spinal-epidural hematoma (SEH)'.)

Perioperative management of anticoagulants for patients who are therapeutically anticoagulated is discussed separately. (See "Perioperative management of patients receiving anticoagulants".)

Duration — In most cases VTE prophylaxis is continued until the patient becomes fully ambulatory or until hospital discharge (typically up to 10 days). Once patients become fully ambulatory, pharmacologic and mechanical methods of prophylaxis are generally stopped. However, the definition of ambulatory is highly subjective such that patients who have prolonged periods of immobility in between ambulatory periods should probably receive continued or additional methods of prophylaxis, particularly in light of the fact that the risk of VTE does not drop to zero upon ambulation but rather may be sustained for weeks beyond the day of surgery [168]. In addition, 10 to 14 days of thromboprophylaxis may be indicated in high-risk surgical patients, including those with cancer. (See "Risk and prevention of venous thromboembolism in adults with cancer", section on 'Surgical patients'.)

Extended pharmacologic VTE prophylaxis beyond discharge is not routinely recommended in most nonorthopedic surgical patients except for those who undergo major abdominal and/or pelvic surgery for cancer. Extended pharmacologic VTE prophylaxis, typically with low molecular weight (LMW) heparin, is offered to this population who are at very high risk for VTE for four weeks postdischarge. The optimal duration of extended prophylaxis is unknown but is typically recommended beyond 10 days and for a period of three to four weeks for high-risk patients who undergo major abdominal and/or pelvic surgery for cancer [12,153-155,169-171]. (See "Risk and prevention of venous thromboembolism in adults with cancer", section on 'Surgical patients'.)

Several meta-analyses of randomized and nonrandomized trials have demonstrated benefit associated with extended prophylaxis in major abdominal/pelvic surgery, particularly for cancer [28,153,155,169,172-174]. As an example, in a meta-analysis of seven randomized trials of patients undergoing major abdominal or pelvic surgery, extended prophylaxis with low molecular weight heparin resulted in a reduction in the overall rate of VTE (5 versus 13 percent) and rate of symptomatic VTE (1 versus 0.1 percent) without an increased risk of bleeding (4 versus 3 percent) [174]. Oral rivaroxaban has been shown to be effective compared with placebo in the extended duration setting among patients who undergo laparoscopic resection for colorectal cancer [175].

Pharmacologic dosing — Pharmacologic prophylaxis is the method of choice for surgical patients at moderate or high risk for VTE (Caprini score ≥3) (table 1). Efficacy is discussed above. (See 'Moderate or high VTE risk' above.)

Low molecular weight heparin — In general, LMW heparin is pharmacologic agent of choice for preventing VTE in nonorthopedic surgical patients. (See 'With low bleeding risk: Pharmacologic alone' above.)

A number of LMW heparin preparations are available (enoxaparin, dalteparin, tinzaparin, nadroparin) none of which have proven superiority over the others when administered as agents to prevent VTE. These regimens differ in detail from product label recommendations but are generally consistent with 2012 American College of Chest Physicians (ACCP) guidelines on antithrombotic therapy and 2013 guidelines of the American Society of Clinical Oncology on VTE prophylaxis and treatment in patients with cancer [12,170].

Typical regimens commonly used in nonorthopedic surgical patients (assuming no renal insufficiency at any point) are listed below. Considerations for individualizing therapy, including adjustments required for neuraxial (ie, epidural or spinal) anesthesia, are reviewed elsewhere. (See 'Timing of initiation' above and 'Duration' above.)

●Enoxaparin:

•Patients without cancer – 40 mg subcutaneously once daily started two hours before abdominal surgery or about 12 hours before other surgery and 40 mg once daily thereafter. Alternatively, 40 mg once daily started within 24 hours after surgery once hemostasis is achieved and initiation is considered safe.

•Patients with cancer – 40 mg 10 to 12 hours before surgery and 40 mg once daily thereafter [170]. Alternatively, 40 mg once daily started approximately 12 to 24 hours after surgery. There is variation in clinical practice especially with regards to the administration of enoxaparin before surgery in patients with cancer. (See "Risk and prevention of venous thromboembolism in adults with cancer", section on 'Surgical patients'.)

•Both once daily and twice daily regimens of enoxaparin have been shown to be effective but direct comparison of these regimens are poorly studied [160]. More aggressive prophylaxis (eg, twice daily enoxaparin and/or the addition of a mechanical device) may be considered in very high-risk populations (eg, patients with cancer).

●Dalteparin:

•Including patients with cancer – 5000 units subcutaneously started about 12 hours (or evening) before surgery and 5000 units once daily thereafter.

•Although labeling in some countries recommends reduced doses in lower thrombotic risk general surgeries (2500 units subcutaneously started 1 to 2 hours before surgery and 2500 units daily thereafter), we use full dosing listed above when LMW prophylaxis is indicated.

Less commonly used agents are:

●Tinzaparin (not available in the United States):

•Including patients with cancer – 4500 units subcutaneously started 12 hours before surgery and once daily thereafter. An alternative weight-based regimen is listed in the drug information monograph.

•Although labeling in some countries recommends reduced doses in lower thrombotic risk general surgeries (3500 units subcutaneously started one to two hours before surgery and 3500 units daily thereafter), we use full dosing listed above when LMW prophylaxis is indicated.

●Nadroparin (not available in the United States):

•High VTE risk including patients with cancer – 38 units/kg (maximum 3800 units) subcutaneously once daily started 12 hours before surgery; on postoperative day 4 increase dose to 57 units/kg (maximum 5700 units once daily).

•Although labeling in some countries recommends reduced doses in lower thrombotic risk general surgeries (2850 units subcutaneously started one to two hours before surgery and 2850 units daily thereafter), we use full dosing listed above when LMW prophylaxis is indicated.

The platelet count should be monitored regularly (eg, day 5, 7, and 9) in all patients receiving LMW heparin to detect the development of heparin-induced thrombocytopenia (HIT). All heparin agents are contraindicated in patients with active HIT or a history of HIT, the details of which are discussed separately. (See "Management of heparin-induced thrombocytopenia".)

The ideal dose for patients with obesity is unknown. In patients with a BMI ≥40 kg/m², some experts empirically increase the standard LMW heparin dose by approximately 30 percent [176]. Suggested doses for patients with obesity and patients undergoing bariatric surgery are presented in the table (table 4) and discussed separately. (See "Bariatric surgery: Postoperative and long-term management", section on 'Venous thromboembolism'.)

Based upon data extrapolated from patients receiving therapeutic doses of LMW heparin, we prefer the avoidance of these agents in patients with severe renal insufficiency (eg, creatinine clearance 20 to 30 mL/min and end stage renal failure requiring dialysis) should it be present upon admission. For those with mild renal insufficiency dose-adjustments can be made according to the creatinine clearance and agent chosen, as outlined in the table (table 2). For those who develop severe renal insufficiency during hospitalization, it is prudent that the LMW heparin agent be discontinued and replaced with unfractionated heparin (UFH). (See "Heparin and LMW heparin: Dosing and adverse effects".)

Low-dose unfractionated heparin — Low-dose UFH is generally considered an alternative to LMW heparin, when, for example, cost or renal insufficiency (creatinine clearance <20 to 30 mL/min) is an issue.

Low-dose subcutaneous UFH for VTE prophylaxis is usually given as 5000 units every 12 hours starting two or more hours before surgery.

Occasionally, if the risk is assessed as particularly high, the frequency is increased to three times daily, although data to support this dosing is lacking. In patients with cancer, 5000 units every eight hours starting two to four hours before surgery is suggested by American Society of Clinical Oncology 2013 guidelines [170]. Factors including cost, institutional policy, body weight, and risk of bleeding may be used to help the clinician make the decision to choose the frequency of dosing.

The ideal dose for patients with obesity is unknown such that dosing should be individualized on a case-by-case basis. We generally prefer to empirically treat with UFH 5000 to 7500 units twice daily; alterations in the dosing and frequency (eg, three times per day dosing) should be tailored to individual patients [177-179].

The dose of UFH does not need to be adjusted for patients with renal insufficiency. However, similar to LMW heparin, the platelet count should be monitored regularly (eg, day 5, 7, and 9) to detect the development of HIT, in which case UFH should be discontinued. (See "Heparin and LMW heparin: Dosing and adverse effects", section on 'Platelet count monitoring'.)

Fondaparinux — Fondaparinux is an alternative to LMW heparin and UFH in patients with contraindications to heparin (eg, HIT) or when these agents are not available.

Unlike LMW heparin agents, which are often administered pre and postoperatively, in accordance with trials that have shown benefit, fondaparinux is typically given as 2.5 mg once daily, starting at least six to eight hours postoperatively (after skin closure). Although in Europe, a 1.5 mg dose is suggested for individuals with creatinine clearance of 20 to 50 mL/min, the safety and efficacy of this dose has not been well-studied. One systematic review of 17 trials reported superior efficacy of fondaparinux compared with LMW heparin (odds ratio [OR] 0.49; 95% CI 0.38-0.64) but at the expense of an increased risk of major bleeding (OR 1.48, 95% CI 1.15-1.9) [180]. Dosing and adverse effects of fondaparinux are discussed separately. (See "Fondaparinux: Dosing and adverse effects".)

METHODS NOT RECOMMENDED

Screening — Secondary prevention with screening tests targeted at the early detection of thrombosis (eg, ultrasonography) is not recommended but can be reserved for rare patients in whom primary prophylaxis is not suitable (eg, patients with active minor bleeding); in such cases, resumption of primary prophylaxis should be performed as soon as is feasible [181,182]. (See "Prevention of venous thromboembolic disease in acutely ill hospitalized medical adults", section on 'Definition of VTE prophylaxis' and "Overview of inpatient management of the adult trauma patient", section on 'Screening for venous thromboembolism'.)

Prophylactic vena cava filters — Inferior vena cava (IVC) filters should generally be avoided as prophylaxis against postoperative VTE. This approach is best supported by indirect evidence from large populations of medical and surgical patients (PREPIC) in whom IVC filters were placed (mostly for therapy) that reported a reduction in pulmonary embolism (PE) but an increase in the rate of lower extremity deep venous thrombosis (DVT). These data are discussed separately. (See "Overview of the treatment of proximal and distal lower extremity deep vein thrombosis (DVT)", section on 'Inferior vena cava filter'.)

Smaller observational studies have confirmed similar outcomes in surgical patients:

●One retrospective study of over 6000 bariatric patients reported that compared with those who did not receive an IVC filter, filter placement did not reduce the rate of VTE and may be associated with an increased risk of death or serious disability [183].

●Systematic reviews and observational studies in trauma patients reported a reduction in the rate of PE but increased rate of DVT and IVC filter complications (eg, migration) without an effect on mortality in those treated with an IVC filter compared with those in whom a filter was not placed [184-186]. In contrast, another randomized study reported no difference in DVT or PE rate in trauma patients with who had an early prophylactic IVC filter placed compared with those who did not have a filter placed [187]. These data are discussed separately. (See "Venous thromboembolism risk and prevention in the severely injured trauma patient".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Superficial vein thrombosis, deep vein thrombosis, and pulmonary embolism".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Beyond the Basics topics (see "Patient education: Deep vein thrombosis (DVT) (Beyond the Basics)" and "Patient education: Warfarin (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Defining the population – Postoperative venous thromboembolism (VTE) is common among patients undergoing nonorthopedic surgery. This population includes patients undergoing general and abdominal-pelvic surgery, bariatric, vascular, plastic/reconstructive, cardiac, and thoracic surgery as well as patients undergoing neurosurgery and patients admitted with major trauma. (See 'Introduction' above.)

●Assess thrombosis risk – The risk of postoperative VTE depends upon procedure-related factors (eg, anatomic location, degree of invasiveness, type and duration of anesthesia, requirement for postoperative immobilization) and patient-related risk factors (eg, increasing age, prior VTE, presence of malignancy) all of which need to be assessed prior to surgery so that a risk category can be assigned and method appropriately selected. (See 'Assess risk for thrombosis' above.)

•For patients undergoing general (eg, breast, thyroid, parathyroid) and abdominal-pelvic (gastrointestinal, urologic, gynecologic) surgery as well as patients undergoing bariatric, vascular, and plastic/reconstructive surgery, the modified Caprini risk assessment score is used to classify patient into very low risk, low risk, moderate risk, and high risk (table 1).

•Patients undergoing major cardiac, thoracic, brain, and spinal surgery or patients with major trauma are at minimum, moderate risk for VTE (eg, uncomplicated cardiac, bariatric, minor thoracic or spinal surgery) and many are at high risk for VTE (eg, craniotomy, extensive cardiac or thoracic surgery, spinal surgery for malignancy, trauma involving brain or spine). (See 'Assess risk for thrombosis' above.)

●Suggested approach based upon VTE risk – In nonorthopedic patients, assuming the risk of bleeding is low, we suggest the following approach (table 1):

•Very low risk – For most patients at very low risk of VTE, pharmacologic or mechanical methods of thromboprophylaxis are typically not necessary since most are fully ambulatory; thus, early and frequent ambulation is the preferred method in this population. (See 'Very low thrombosis risk: Early ambulation' above.)

•Low risk – For patients at low risk of VTE, we suggest mechanical methods of VTE prophylaxis, rather than pharmacologic or no thromboprophylaxis (Grade 2C). Choosing among mechanical methods is individualized, but intermittent pneumatic compression (IPC) devices or graduated compression stockings (GCS) are frequently chosen. (See 'Low VTE risk: Mechanical methods' above.)

•Moderate risk – For patients at moderate risk of VTE, we suggest pharmacologic prophylaxis, preferably with low molecular weight (LMW) heparin, rather than no thromboprophylaxis (Grade 2B). (See 'Moderate or high VTE risk' above and 'Pharmacologic dosing' above.)

•High risk – For patients at high risk of VTE, we recommend, at minimum, pharmacologic prophylaxis, preferably with LMW heparin, rather than mechanical methods or no prophylaxis (Grade 1B). Combining pharmacologic with mechanical methods of thromboprophylaxis is frequently performed in this population especially in those at greatest risk (eg, multiple risk factors or cancer). (See 'Moderate or high VTE risk' above.)

•High bleeding risk – For nonorthopedic surgical patients with contraindications to pharmacologic prophylaxis and in whom thromboprophylaxis other than ambulation is indicated (table 3), patients with active or at high risk of bleeding or patients in whom the consequences of bleeding are thought to be potentially catastrophic (eg, neurosurgical procedures), mechanical methods rather than no thromboprophylaxis are preferred. IPC devices are commonly used, but GCS or venous foot pump devices are also appropriate (picture 1 and picture 2). Switching to or adding a pharmacologic agent, such as LMW heparin, should be done as soon as the bleeding risk becomes acceptably low (eg, 48 to 72 hours following neurosurgery) or the bleeding diathesis has been reversed.

•Renal insufficiency – For most patients without renal insufficiency (eg, creatinine clearance >30 mL/min) in whom pharmacologic prophylaxis is indicated, we suggest LMW heparin rather than unfractionated heparin (UFH) (table 2) (Grade 2C). UFH is preferred in those with severe renal insufficiency (eg, creatinine clearance <20 to 30 mL/min), while fondaparinux is preferred in those with heparin-induced thrombocytopenia. Oral agents including warfarin, aspirin, and direct oral anticoagulants are unstudied and are not typically administered in nonorthopedic surgical patients. (See 'With low bleeding risk: Pharmacologic alone' above.)

●Timing – The optimal timing for initiation of thromboprophylaxis is unknown and should be individualized.

•For most patients in whom the bleeding risk is low, mechanical methods are ideally started just before surgery and pharmacologic agents administered within 2 to 12 hours before surgery, with the exception of fondaparinux, which is typically initiated 6 to 8 hours after skin closure. Postoperatively, thromboprophylaxis is started within 24 hours of surgery, typically on the evening of the day of surgery (6 to 8 hours postoperatively) or on the morning of the day after surgery. (See 'Timing of initiation' above.)

•However, in patients with a contraindication to anticoagulants, a high risk of bleeding, or potential catastrophic effect of bleeding, mechanical methods should be employed postoperatively and pharmacologic agents started or added postoperatively, when adequate hemostasis is achieved and it is assessed as safe (eg, after 48 to 72 hours).

●Duration – In most patients, we suggest that VTE prophylaxis be continued until the patient becomes fully ambulatory or until hospital discharge. Extended pharmacologic VTE prophylaxis beyond discharge is not routinely recommended in this population except for patients who undergo major abdominal and/or pelvic surgery for cancer in whom thromboprophylaxis should be continued for a period of four weeks. (See 'Duration' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Menaka Pai, MD, FRCPC, who contributed to earlier versions of this topic review.