American Society of Hematology 2019 guidelines for management of venous thromboembolism: prevention of venous thromboembolism in surgical hospitalized patients

These American Society of Hematology (ASH) guidelines are based on updated and original systematic reviews of evidence conducted by researchers and developed under the direction of the McMaster University GRADE Centre with international collaborators. The panel followed best practice for guideline development recommended by the Institute of Medicine (now the National Academy of Medicine) and the Guidelines International Network.^1-4  The panel used the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach^5,6  to assess the certainty in the evidence and formulate recommendations.

The population of postoperative patients is heterogeneous with regard to the degree of risk of venous thromboembolism (VTE), depending on intrinsic patient factors and those factors that are related to the type of surgery, mobilization, anatomic location of the procedures, and risk of bleeding. Surgeons have a long history of accepting prophylactic measures against VTE, be they mechanical or pharmacological.

The strength of a recommendation is expressed as strong (“the guideline panel recommends…”) or conditional (“the guideline panel suggests…”) and has the following interpretation:

• For patients: most individuals in this situation would want the recommended course of action, and only a small proportion would not.

• For clinicians: most individuals should follow the recommended course of action. Formal decision aids are not likely to be needed to help individual patients make decisions consistent with their values and preferences.

• For policy makers: the recommendation can be adopted as policy in most situations. Adherence to this recommendation according to the guideline could be used as a quality criterion or performance indicator.

• For researchers: the recommendation is supported by credible research or other convincing judgments that make additional research unlikely to alter the recommendation. On occasion, a strong recommendation is based on low or very low certainty in the evidence. In such instances, further research may provide important information that alters the recommendation.

• For patients: the majority of individuals in this situation would want the suggested course of action, but many would not. Decision aids may be useful in helping patients to make decisions consistent with their individual risks, values, and preferences.

• For clinicians: different choices will be appropriate for individual patients, and clinicians must help each patient to arrive at a management decision consistent with the patient’s values and preferences. Decision aids may be useful in helping individuals to make decisions consistent with their individual risks, values, and preferences.

• For policy makers: policy-making will require substantial debate and involvement of various stakeholders. Performance measures about the suggested course of action should focus on whether an appropriate decision-making process is duly documented.

• For researchers: this recommendation is likely to be strengthened (for future updates or adaptation) by additional research. An evaluation of the conditions and criteria (and the related judgments, research evidence, and additional considerations) that determined the conditional (rather than strong) recommendation will help to identify possible research gaps.

For patients undergoing major surgery, the ASH guideline panel suggests the following:

1. Using pharmacological prophylaxis or mechanical prophylaxis (conditional recommendation based on low certainty in the evidence of effects ⊕⊕◯◯)

2. For patients who do not receive pharmacologic prophylaxis, using mechanical prophylaxis over no mechanical prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯).

3. For patients who receive mechanical prophylaxis, using intermittent compression devices over graduated compression stockings (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯).

4. For patients who receive pharmacologic prophylaxis, using combined prophylaxis with mechanical and pharmacological methods over prophylaxis with pharmacological agents alone (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯).

5. Depending on the risk of VTE and bleeding based on the individual patient and the type of surgical procedure, using combined prophylaxis or mechanical prophylaxis alone (conditional recommendation based on low certainty in the evidence of effects ⊕⊕◯◯).

Remarks: For patients considered at high risk of bleeding, the balance of effects may favor mechanical methods over pharmacological prophylaxis. For patients considered at high risk for VTE, combined prophylaxis is particularly favored over mechanical or pharmacological prophylaxis alone.

For patients undergoing major surgery, the ASH guideline panel suggests against using inferior vena cava (IVC) filters for prophylaxis of VTE (conditional recommendation based on very low certainty in the evidence of effects, ⊕◯◯◯).

For patients undergoing major surgery, the ASH guideline panel suggests using extended antithrombotic prophylaxis over short-term antithrombotic prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). The ASH guideline panel further suggests using early or delayed antithrombotic prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). Remarks: Extended prophylaxis was generally considered as beyond 3 weeks (range, 19-42 days) compared with short-term prophylaxis, which was considered as up to 2 weeks (range, 4-14 days). Twelve hours following surgery was arbitrarily selected to be the cutoff point between early and late postoperative antithrombotic administration.

For patients undergoing total hip arthroplasty or total knee arthroplasty, the ASH guideline panel suggests using aspirin (ASA) or anticoagulants (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). When anticoagulants are used, the panel suggests using direct oral anticoagulants (DOACs) over low-molecular-weight heparin (LMWH) (conditional recommendation based on moderate certainty in the evidence of effects ⊕⊕⊕◯); the panel suggests using any of the DOACs approved for use (conditional recommendation based on low certainty in the evidence of effects ⊕⊕◯◯). If a DOAC is not used, the panel suggests using LMWH rather than warfarin (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯) and recommends LMWH rather than unfractionated heparin (UFH) (strong recommendation based on moderate certainty in the evidence of effects ⊕⊕⊕◯).

For patients undergoing hip fracture repair, the ASH guideline panel suggests using pharmacological prophylaxis over no pharmacological prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯) and suggests using LMWH or UFH (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯).

For patients undergoing major general surgery, the ASH guideline panel suggests using pharmacological prophylaxis over no pharmacological prophylaxis (conditional recommendation based on low certainty in the evidence of effects ⊕⊕◯◯) and suggests using LMWH or UFH (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯).

For patients undergoing laparoscopic cholecystectomy, the ASH guideline panel suggests against using pharmacological prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). Remark: Patients with other risk factors for VTE (such as history of VTE, thrombophilia, or malignancy) may benefit from pharmacological prophylaxis.

For patients undergoing major neurosurgical procedures, the ASH guideline panel suggests against using pharmacological prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). For the subset of patients undergoing major neurosurgical procedures for whom pharmacological prophylaxis is used, the ASH guideline panel suggests using LMWH over UFH (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). Remarks: Patients undergoing major neurosurgical procedures are expected to receive prophylaxis with mechanical methods. Pharmacological prophylaxis may be warranted in a higher-risk subgroup of patients, such as those experiencing prolonged immobility following surgery. In addition, pharmacological prophylaxis could be considered for patients undergoing major neurosurgical procedures that carried a lower risk for major bleeding and in those patients with persistent mobility restrictions after the bleeding risk declines following surgery.

For patients undergoing transurethral resection of the prostate (TURP), the ASH guideline panel suggests against using pharmacological prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). For the subset of patients undergoing TURP for whom pharmacological prophylaxis is used, the ASH guideline panel suggests using LMWH or UFH (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). Remark: Patients with other risk factors for VTE (such as history of VTE, thrombophilia, or malignancy) may benefit from pharmacological prophylaxis.

For patients undergoing radical prostatectomy, the ASH guideline panel suggests against using pharmacological prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). For patients undergoing radical prostatectomy in whom pharmacological prophylaxis is used, the ASH guideline panel suggests using LMWH or UFH (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). Remark: Patients undergoing an extended node dissection and/or open radical prostatectomy may have a higher VTE risk and may potentially benefit from pharmacological prophylaxis.

For patients undergoing cardiac or major vascular surgery, the ASH guideline panel suggests using pharmacological prophylaxis or no pharmacological prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯). When pharmacological prophylaxis is used, the panel suggests using LMWH or UFH (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯).

For patients experiencing major trauma and who are at low to moderate risk for bleeding, the ASH guideline panel suggests using pharmacological prophylaxis over no pharmacological prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯).

For patients experiencing major trauma and who are at high risk for bleeding, the ASH guideline panel suggests against pharmacological prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯).

For patients experiencing major trauma in whom pharmacological prophylaxis is used, the ASH guideline panel suggests using LMWH or UFH (conditional recommendation based on low certainty in the evidence of effects ⊕⊕◯◯).

For patients undergoing major gynecological surgery, the ASH guideline panel suggests using pharmacological prophylaxis over no pharmacological prophylaxis (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯) and suggests using LMWH or UFH (conditional recommendation based on very low certainty in the evidence of effects ⊕◯◯◯).

The purpose of these guidelines is to provide evidence-based recommendations about the prevention of VTE for patients undergoing major surgical procedures. The target audience includes patients, surgeons, intensivists, internists, hematologists, general practitioners, hospitalists, other clinicians, pharmacists, and decision makers. Policy makers interested in these guidelines include those involved in developing local, national, or international programs aiming to safely reduce the incidence of VTE and/or to evaluate direct and indirect harms and costs related to VTE and its prevention. This document may also serve as the basis for adaptation by local, regional, or national guideline panels.

Deep vein thrombosis (DVT) and pulmonary embolism (PE) (collectively, VTE) are well-recognized, clinically important, and potentially devastating complications that may occur following major surgical procedures, defined as any surgical intervention that carries greater than minimal risk, is performed in the operating room, and requires specialized training. Before the era of the routine use of effective prophylaxis, VTE was a common cause of morbidity and mortality following major surgery. It has been estimated to cause >50 000 deaths per annum in the United States alone.⁷  The importance of preventative measures to minimize the risk of VTE following major surgery has been recognized for decades; however, even with the use of prophylaxis, surgery accounts for ∼25% of VTEs observed in communities.⁸ 

Although most surgical procedures carry some risk for VTE, this risk varies considerably across surgical procures and among individual patients undergoing surgery. Surgical procedures carrying the highest risk of developing postoperative VTE include hip and knee arthroplasty, invasive neurosurgical procures, and major vascular procedures.⁹ 

Patient factors that carry greater risks for thrombosis include histories of VTE, particularly if unprovoked or associated with cancer, or cancer, even in the absence of previous VTE. Scoring systems that calculate the risk of postoperative VTE for individual patients, such as the Caprini score, have been developed and validated following some surgical procedures.¹⁰ 

Although postoperative VTE has historically been a complication primarily occurring in the hospital, with shortened hospital stays, postoperative VTE often occurs in the days to weeks following discharge from the hospital.¹¹ 

The primary target population of this guideline is patients hospitalized for major surgical procedures that carry a risk for postoperative VTE. This guideline also addresses patients hospitalized following major trauma; most, but not all, subsequently required major surgical procedures. The panel recognized that there are 2 major modalities applied for the prevention of VTE in the postoperative period: pharmacological antithrombotic prophylaxis and mechanical prophylaxis. For evaluation of the pharmacological methods for the prevention of VTE, the panel weighed the benefits and risks of the various options for individual surgical procedures or domains, such as hip or knee arthroplasty, general surgery, or urological or neurosurgical procedures.

For mechanical interventions, such as graduated or mechanical compression devices or IVC filters, the effectiveness of these interventions was assessed across all surgical domains. Likewise, other questions, such as the duration of pharmacological prophylaxis and timing of the initiation of pharmacological prophylaxis, were also assessed across all surgical domains.

The target populations included patients who underwent surgery for cancer or noncancer-related procedures. Patients hospitalized for major trauma were included whether they underwent surgery or not.

The guideline panel developed and graded the recommendations and assessed the certainty in the supporting evidence following the GRADE approach.^5,6,12-16  The overall guideline-development process, including funding of the work, panel formation, management of conflicts of interest, internal and external review, and organizational approval, was guided by ASH policies and procedures derived from the Guidelines International Network–McMaster Guideline Development Checklist (http://cebgrade.mcmaster.ca/guidecheck.html) and was intended to meet recommendations for trustworthy guidelines by the Institute of Medicine and the Guidelines International Network.^1-4 

The work of this panel was coordinated with 9 other guideline panels (addressing other aspects of VTE management) by ASH and the McMaster GRADE Centre (funded by ASH under a paid agreement). Project oversight was provided initially by a coordination panel, which reported to the ASH Committee on Quality, and then by the coordination panel chair (Adam Cuker) and vice chair (H.J.S.). ASH vetted and appointed individuals to the guideline panel. The McMaster GRADE Centre vetted and retained researchers to conduct systematic reviews of evidence and coordinate the guideline-development process, including the use of the GRADE approach. The membership of the panels and the GRADE Centre team is described in Supplement 1.

The panel included surgeons with subspecialty representation, hematologists, internists, and a pharmacist, all of whom had clinical and research expertise on the guideline topic. The panel also included methodologists with expertise in evidence appraisal and guideline development and 2 patient representatives. Both patient representatives participated in question prioritization, and 1 participated in all remaining steps of the development process. The panel chair was a content expert. The vice chair was a urological surgeon with specialized expertise in guideline development.

In addition to synthesizing evidence systematically, the McMaster GRADE Centre supported the guideline-development process, including determining methods, preparing agendas and meeting materials, and facilitating panel discussions. The panel’s work was done using Web-based tools (www.surveymonkey.com and www.gradepro.org) and face-to-face and online meetings.

Development of these guidelines was wholly funded by ASH, a nonprofit medical specialty society that represents hematologists. Some members of the guideline panel were members of ASH. ASH staff supported panel appointments and coordinated meetings but had no role in choosing the guideline questions or determining the recommendations.

Members of the guideline panel received travel reimbursement for attendance at in-person meetings. The patient representative (C.B.) received an honorarium of $200. The panelists received no other payments. Some researchers who contributed to the systematic evidence reviews received salary or grant support through the McMaster GRADE Centre. Other researchers participated to fulfill requirements of an academic degree or program.

Conflicts of interest of all participants were managed according to ASH policies based on recommendations of the Institute of Medicine¹⁷  and the Guidelines International Network.⁴  At the time of appointment, a majority of the guideline panel, including the chair and the vice chair, had no conflicts of interest as defined and judged by ASH (ie, no current material interest in any commercial entity with a product that could be affected by the guidelines). Some panelists disclosed new interests or relationships during the development process, but the majority continued to have no conflicts of interest with commercial entities, as judged by ASH.

Before appointment to the panel, individuals disclosed financial and nonfinancial interests. Members of the VTE Guideline Coordination Panel reviewed the disclosures and judged which interests were conflicts and should be managed. Supplement 2 provides the complete “Disclosure of Interests” forms of all panel members. In Part A of the forms, individuals disclosed material interests for 2 years prior to appointment. In Part B, they disclosed other interests that were not mainly financial. Part C summarizes ASH decisions about which interests were judged to be conflicts. Part D describes new interests disclosed by individuals after appointment.

Recusal was used to manage conflicts of interest. During deliberations, panel members with a current direct financial interest in a commercial entity with any product that could be affected by the guidelines participated in discussions about the evidence and clinical context but were recused from making judgments or voting about individual domains (eg, magnitude of desirable consequences) and the direction and strength of relevant recommendations.^4,18-20  The Evidence-to-Decision (EtD) framework for each recommendation describes which individuals were recused from making judgments about each recommendation.

None of the McMaster-affiliated researchers who contributed to the systematic evidence reviews or who supported the guideline-development process had any current material interest in a commercial entity with any product that could be affected by the guidelines. Supplement 3 provides the complete “Disclosure of Interest” forms of researchers who contributed to these guidelines.

The panel used the GRADEpro Guideline Development Tool (www.gradepro.org) and SurveyMonkey (www.surveymonkey.com) to brainstorm and then prioritize the questions described in Table 1.

The panel selected outcomes of interest for each question a priori, following an approach described in detail elsewhere.²¹  The panel rated the following outcomes as critical for clinical decision making across all questions: mortality, symptomatic PEs, symptomatic proximal DVTs, symptomatic severe distal DVTs, major bleeding, and reoperation. The panel used an explicit process to rate the clinical severity of DVTs and PEs. The panel determined that all symptomatic proximal DVTs and PEs confirmed by objective diagnostic imaging were considered of moderate severity and were clinically important.²²  For several outcomes, the studies reported outcomes that were different or were in addition to the outcomes that the panel determined to be important for decision making. Typically, included studies reported outcomes as any PE, any DVT, or any proximal or distal DVT. Some studies did not distinguish asymptomatic thromboembolic events that were detected by the routine performance of sensitive screening studies for VTEs from symptomatic thromboembolic events where patients developed overt symptoms that were subsequently confirmed by objective testing to be associated with VTEs. Reporting of symptomatic thromboembolic events was inconsistent across studies.

Where available, questions were addressed with studies that reported symptomatic outcome events. In the absence of reports of symptomatic VTE in a clinical question, modeling was performed using estimates of the proportion of asymptomatic DVTs that would become clinically important symptomatic events (Supplement 4). Only “severe” distal DVTs were rated as clinically important distal DVTs, and it was estimated that only ∼25% of symptomatic postoperative distal DVTs would be considered severe.

Major bleeding definitions varied across clinical studies. For the purposes of this analysis, outcome events that met the definition of major bleeding for individual studies were applied. The exception was that the need for a blood transfusion itself was not considered major bleeding unless other criteria for major bleeding were met. The definition of reoperation was not specific for reoperation caused by or related to major bleeding.

Studies evaluated included patients with cancer and without cancer. We tested potential differences in the effects on studies with >50% and <50% of participants with cancer. Subgroup analyses did not demonstrate a difference in the relative effectiveness of interventions whether cancer patients were included or not. As a result, recommendations do not distinguish between cancer and noncancer patients.

For the use of pharmacological and mechanical methods of prophylaxis, the panel advises to follow manufacturer’s recommendations regarding patient-specific restrictions in the use of individual products (such as levels of renal function for patients receiving LMWHs or DOACs). The panel also advises periodic monitoring of the platelet count for patients receiving LMWH and, in particular, UFH, as postoperative prophylaxis in consideration of the risk of heparin-induced thrombocytopenia.

For each guideline question, the McMaster GRADE Centre prepared a GRADE EtD framework, using the GRADEpro Guideline Development Tool (www.gradepro.org).^12,13,16  The EtD table summarized the results of systematic reviews of the literature that were updated or performed for this guideline. The EtD table addressed effects of interventions, resource utilization (cost-effectiveness), values and preferences (relative importance of outcomes), equity, acceptability, and feasibility. The guideline panel reviewed draft EtD tables before, during, or after the guideline panel meeting and made suggestions for corrections and identified missing evidence. To ensure that recent studies were not missed, searches (Supplement 5) were updated during October and November 2016, and panel members were asked to suggest any studies that may have been considered missed and fulfilled the inclusion criteria for the individual questions. Monthly search alerts were created and monitored to capture relevant new studies up to 1 July 2019, prior to submission of the manuscript for publication.

Under the direction of the McMaster GRADE Centre, researchers followed the general methods outlined in the Cochrane Handbook for Systematic Reviews of Interventions (handbook.cochrane.org) for conducting updated or new systematic reviews of intervention effects. When existing reviews were used, judgments of the original authors about risk of bias were randomly checked for accuracy and accepted or conducted de novo if they were not available or not reproducible. For new reviews, risk of bias was assessed at the health outcome level using the Cochrane Collaboration’s risk of bias tool for randomized trials or nonrandomized studies. In addition to conducting systematic reviews of intervention effects, the researchers searched for evidence related to baseline risks, values, preferences, and costs and summarized findings within the EtD frameworks.^12,13,16  Subsequently, the certainty in the body of evidence (also known as quality of the evidence or confidence in the estimated effects) was assessed for each effect estimate of the outcomes of interest following the GRADE approach based on the following domains: risk of bias, precision, consistency and magnitude of the estimates of effects, directness of the evidence, risk of publication bias, presence of large effects, dose-response relationship, and an assessment of the effect of residual, opposing confounding. The certainty was categorized into 4 levels ranging from very low to high.^5,6,14 

During a 2-day in-person meeting, followed by online communication and conference calls, the panel developed clinical recommendations based on the evidence summarized in the EtD tables. For each recommendation, the panel took a population perspective and came to consensus on the following: the certainty in the evidence, the balance of benefits and harms of the compared management options, and the assumptions about the values and preferences associated with the decision. The guideline panel also explicitly took into account the extent of resource use associated with alternative management options. The panel agreed on the recommendations (including direction and strength), remarks, and qualifications by consensus or, in rare instances, by voting (an 80% majority was required for a strong recommendation), based on the balance of all desirable and undesirable consequences. In the event that consensus was not reached based on discussion, the recommendation was made based on a vote of the panel, with the results of such votes listed in the text following the recommendation. The final guidelines, including recommendations, were reviewed and approved by all members of the panel.

The recommendations are labeled as “strong” or “conditional” according to the GRADE approach. The words “the guideline panel recommends” are used for strong recommendations, and “the guideline panel suggests” is used for conditional recommendations. Table 2 provides the suggested interpretation of strong and conditional recommendations by patients, clinicians, and health care policy makers.

Draft recommendations were reviewed by all members of the panel, revised, and then made available online on 22 June 2018 for external review by stakeholders, including allied organizations, other medical professionals, patients, and the public. Sixteen individuals or organizations submitted comments. The document was revised to address pertinent comments, but no changes were made to the recommendations. The guidelines were reviewed by the ASH Guideline Oversight Subcommittee on 28 August 2019, approved by the Committee on Quality on 6 September 2019 and by the ASH officers on 13 September 2019, and then subjected to peer review.

ASH guidelines are primarily intended to help clinicians make decisions about diagnostic and treatment alternatives. Other purposes are to inform policy, education, and advocacy and to state future research needs. They may also be used by patients. These guidelines are not intended to serve or be construed as a standard of care. Clinicians must make decisions on the basis of the clinical presentation of each individual patient, ideally through a shared process that considers the patient’s values and preferences with respect to the anticipated outcomes of the chosen option. Decisions may be constrained by the realities of a specific clinical setting and local resources, including, but not limited to, institutional policies, time limitations, and availability of treatments. These guidelines may not include all appropriate methods of care for the clinical scenarios described. As science advances and new evidence becomes available, recommendations may become outdated. Following these guidelines cannot guarantee successful outcomes. ASH does not warrant or guarantee any products described in these guidelines.

Statements about the underlying values and preferences, as well as qualifying remarks accompanying each recommendation, are its integral parts and serve to facilitate more accurate interpretation. They should never be omitted when recommendations from these guidelines are quoted or translated. Implementation of the guidelines will be facilitated by the related interactive forthcoming decision aids. The use of these guidelines is also facilitated by the links to the EtD frameworks and interactive summary-of-findings tables in each section.

Question: Should pharmacological prophylaxis vs mechanical prophylaxis be used for patients undergoing major surgery?

We identified 11 systematic reviews addressing, in part, this question.^23-33  We identified 38 studies in this review that fulfilled our inclusion criteria and measured outcomes relevant to this context.^34-72  Our systematic search of randomized controlled trials (RCTs) identified 2 additional studies not included in previous systematic reviews and that fulfilled the inclusion criteria.^70,72 

Fifteen studies reported the effect of the pharmacological prophylaxis compared with mechanical prophylaxis alone on risk of mortality.^{35,36,40,42,51,54,57,59-61,63-65,68,72}  Thirteen studies reported the effect on the development of symptomatic PEs,^{35,37,39,43,45,62,64,67-72}  and 17 studies reported the effect on the development on any PE.^{34,36,38,40,42,46,47,50,51,53,58,60,61,63,67,71}  Six studies reported data on symptomatic DVTs,^35,67,69-72  and 17 studies reported data on any proximal DVT.^{34,35,37,39,41,45,52,53,57,58,64,65,67,69-72}  Four studies reported data on symptomatic distal DVTs,^35,70-72  and 16 studies reported data on any distal DVT.^{34,35,37,39,40,45,52,53,57,58,63,64,66,67,71,72} 

Eighteen studies reported the effect of pharmacological prophylaxis compared with mechanical prophylaxis alone on the risk of major bleeding,^{35-37,41,42,43,48,51,54,59-61,63-66,68,72,} and 6 studies reported the effect on the risk of reoperation.^{37,42,46,48,54,72} 

The EtD framework is available at https://guidelines.gradepro.org/profile/B57A59FE-FCA9-8C9A-8C8D-55089B4E8FB1.

The systematic review found that there may be no difference in mortality between pharmacological and mechanical prophylaxis (relative risk [RR], 0.92; 95% confidence interval [CI], 0.46-1.84; low certainty in the evidence of effects); this corresponds to 1 fewer (5 fewer to 7 more) death per 1000 patients. Similarly, irrespective of the baseline risk chosen, which was derived from a cohort study of 172 320 patients,⁷³  there may be no difference for symptomatic PEs (RR, 1.04; 95% CI, 0.36-2.96; low certainty in the evidence of effects), corresponding to 0 fewer events (2 fewer to 7 more). There is also likely little to no difference in symptomatic proximal DVTs (RR, 0.75; 95% CI, 0.11-5.32; moderate certainty in the evidence of effects); depending on the baseline risk of 1.6% or 2.6%,⁷³  this corresponds to anywhere from 4 fewer (15 fewer to 71 more) to 7 fewer (23 fewer to 113 more) events, respectively. For symptomatic distal DVTs, pharmacological prophylaxis likely results in a reduction in risk (RR, 0.16; 95% CI, 0.05-0.58; moderate certainty in the evidence of effects); however, this corresponds to a possibly small, and likely unimportant, reduction in symptomatic distal DVTs in absolute terms of 2 fewer (1-2 fewer) per 1000 patients, based on a baseline risk of 1.2% from observational data.⁷³ 

Pharmacological prophylaxis likely leads to more major bleeding (RR, 2.87; 95% CI, 1.68-4.92; moderate certainty in the evidence of effects). This corresponds to a small absolute increase of 12 more (4-25 more) major bleeds per 1000 patients. Pharmacological prophylaxis probably results in no difference in reoperations (RR, 2.01; 95% CI, 0.29-14.05; low certainty in the evidence of effects), corresponding to 1 more (1 fewer to 19 more) per 1000 patients.

We rated the overall certainty in the evidence of effects as low based on the lowest certainty in the evidence for the critical outcomes, downgrading twice for very serious imprecision.

The panel determined that, on balance, neither approach was favored over the other for patients undergoing major surgery at low or moderate risk for major bleeding because of the low certainty in the evidence, as well as concerns about compliance with mechanical prophylaxis. For patients at high baseline risk for major bleeding, mechanical prophylaxis would more clearly be favored because of the incremental risk of bleeding with pharmacological prophylaxis. There was possibly important uncertainty or variability in how much affected individuals valued the main outcomes. The panel further judged the costs associated with pharmacological prophylaxis to be moderate based on very low certainty in the evidence about resource requirements. Cost-effectiveness probably favors mechanical prophylaxis. The panel did not account for potential risks of mechanical prophylaxis, including fall risk, risk of skin damage, and limitation in mobility. There would probably be no impact on health equity, and pharmacological and mechanical prophylaxis would be acceptable to stakeholders and probably feasible to implement. The panel recognized that most of the evidence informing this recommendation came from the orthopedic literature (elective knee and hip arthroplasty).

The guideline panel suggests using pharmacological prophylaxis or mechanical prophylaxis for patients undergoing major surgery, based on low certainty in the evidence of effects. For patients deemed at high risk for major bleeding (because of the nature of the surgical procedure), there is likely a net benefit in favor of mechanical prophylaxis.

The panel determined that it would be valuable to have further high-quality studies comparing these interventions outside of the orthopedic setting to confirm the generalizability of the results across surgical domains. The panel would also welcome high-quality studies to determine the effectiveness of mechanical prophylaxis administered outside the hospital setting. The panel identified the need for more and better studies on how patients value the various outcomes in the perioperative setting and to what degrees these values vary by patients as a future research priority.

Question: Should mechanical prophylaxis vs no prophylaxis be used for patients undergoing major surgery?

We identified 5 systematic reviews addressing, in part, this question.^{24,25,27,28,30}  We identified 25 studies in these reviews that fulfilled our inclusion criteria and measured outcomes relevant to this context.^{37,49,53,66,74-94}  Our systematic search of RCTs did not identify any additional study that fulfilled the inclusion criteria.

Ten studies reported the effect of mechanical prophylaxis compared with no prophylaxis on risk of mortality.^{49,77,78,81,83-85,88,90,94}  Nine studies reported the effect on the development of symptomatic PEs,^{37,76,78,81,83,84,86,88,92}  and 5 studies reported the effect on any PE.^{53,77,85,86,93}  No study reported data on symptomatic proximal or distal DVT, but 8 studies reported on any proximal DVT,^{37,78,79,82,84,85,92,93}  and 7 studies reported on any distal DVT.^{37,77,79,82,85,92,93} 

The EtD framework is available online at https://guidelines.gradepro.org/profile/61E7ADC1-4C91-8D58-9E3A-56BFEE3EAC20.

There may be no difference in mortality between mechanical prophylaxis and no prophylaxis (RR, 1.33; 95% CI, 0.71-2.51; low certainty in the evidence of effects); this corresponds to 6 more (5 fewer to 28 more) deaths per 1000 patients. There may be a small difference in symptomatic PEs (RR, 0.61; 95% CI, 0.27-1.40; low certainty in the evidence of effects) corresponding to 3 fewer (6 fewer to 3 more) symptomatic PEs per 1000 patients based on a baseline risk of 0.8% and 4 fewer (8 fewer to 4 more) per 1000 patients based on a baseline risk of 1.1% from observational data.⁷³  There may be no difference in symptomatic proximal DVTs (RR, 0.75; 95% CI, 0.35-1.61; very low certainty in the evidence of effects), but we are uncertain of this. The risk of symptomatic distal DVTs may be reduced (RR, 0.66; 95% CI, 0.50-0.86; very low certainty in the evidence of effects), but we are uncertain of this.

There were no relevant adverse events deemed critical for this comparison. The panel was unable to assess the relative effect of mechanical prophylaxis on potential hazards, such as falls or skin complications.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, imprecision, and inconsistency.

The panel determined that there was possibly important uncertainty or variability in how much affected individuals valued the main outcomes. They further judged that the balance between desirable and undesirable effects probably favors mechanical prophylaxis over no prophylaxis for patients at moderate or high risk for VTE but not patients at low risk for VTE. The panel judged the costs associated with mechanical prophylaxis to be moderate based on very low certainty in the evidence about resource requirements, with no available studies explicitly addressing this question. Cost-effectiveness probably favors mechanical prophylaxis. There would probably be no impact on health equity; mechanical prophylaxis would probably be acceptable to stakeholders and probably feasible to implement. The panel recognized that many patients considered at moderate to high risk for VTE would receive pharmacological prophylaxis in addition to mechanical methods. The guideline panel determined that there was very low certainty evidence for a net health benefit/harm for mechanical prophylaxis. Most of the evidence comes from orthopedics (elective hip and knee arthroplasty). In settings where intermittent pneumatic compression is not available, the use of stockings as mechanical prophylaxis is an acceptable and feasible option (see Recommendation 3).

For patients undergoing major surgery and at risk for VTE, the ASH guideline panel suggests using mechanical prophylaxis over no mechanical prophylaxis, recognizing that the certainty in the evidence is very low for this recommendation.

The panel recognizes that there is a need for high-quality clinical trials using clinically relevant end points to improve the certainty of the evidence supporting this recommendation, particularly outside the orthopedic setting. However, this is likely a lower priority for research than studies evaluating mechanical prophylaxis in combination with pharmacological prophylaxis.

Question: Should pneumatic compression devices vs graduated compression stockings be used for patients undergoing major surgery?

We identified 6 systematic reviews addressing this question.^25-28,31,32  We identified 11 studies in these reviews that fulfilled our inclusion criteria and measured outcomes relevant to this context.^37,94-103  Our systematic search of RCTs did not identify any additional study that fulfilled the inclusion criteria.

Five studies reported the effect of pneumatic compression prophylaxis compared with graduated compression stockings prophylaxis on risk of mortality.^{94,96,97,101,102}  Eight studies reported the effect on the development of symptomatic PEs,^{37,95,96,97,98,99,102,103}  and 4 studies reported the effect on any PE.^{94,100,101,103}  One study reported data on symptomatic proximal and symptomatic distal DVTs,⁹⁸  whereas 6 studies reported on any proximal DVT,^{37,94,96,98-100}  and 5 studies reported on any distal DVT.^{37,94,96,98,100} 

The EtD framework is available online at https://guidelines.gradepro.org/profile/1584FD2F-9CC6-9C59-8DF5-48F0045F1BE5.

There is no difference in mortality between pneumatic compression and graduated compression stockings prophylaxis (RR, 1.04; 95% CI, 0.16-6.63; low certainty in the evidence of effects); this corresponds to 2 more (41 fewer to 274 more) per 1000 patients. There may also be no difference in symptomatic PEs (RR, 0.56; 95% CI, 0.17-1.86; low certainty in the evidence of effects). The risk of symptomatic proximal DVTs may be reduced (RR, 0.48; 95% CI, 0.25-0.93, very low certainty in the evidence of effects), but we are uncertain of this finding. This benefit likely corresponds to 9 fewer (1-12 fewer) symptomatic proximal DVTs in 1000 patients with a baseline risk of 1.6% and 14 fewer (2-20 fewer) symptomatic proximal DVTs per 1000 patients based on a baseline risk of 2.6% from observational data.⁷³  The risk of distal DVT (RR, 0.55; 95% CI, 0.25-1.22, very low certainty in the evidence of effects) appears to be similar, but we are uncertain of this finding.

There were no relevant adverse events deemed critical for this comparison. Potential harms included reduced mobility, and pneumatic compression prophylaxis may be uncomfortable. There is also a small risk for inappropriate use of pneumatic compression prophylaxis for some patients (eg, those with lower extremity fractures).

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, imprecision, and inconsistency.

The panel determined that there was possibly important uncertainty or variability in how much affected individuals valued the main outcomes. They further judged that the balance between desirable and undesirable effects probably favors pneumatic compression prophylaxis over graduated compression stockings prophylaxis. The panel was unable to assess the relative effect of pneumatic compression compared with that of graduated compression stockings on the risk of other hazards, such as falls or skin complications. These might be considered “unmeasured harms” of mechanical prophylaxis. The panel judged the costs associated with pneumatic compression prophylaxis to be moderate based on very low certainty in the evidence about resource requirements, with no available studies explicitly addressing this question. Cost-effectiveness probably favors pneumatic compression prophylaxis. There would probably be no impact on health equity; pneumatic compression prophylaxis would probably be acceptable to stakeholders and probably feasible to implement. Lack of information regarding out-of-hospital use of pneumatic compression was a limitation of this technique. The panel recognizes that most of the evidence about the effectiveness comes from orthopedics (elective hip and knee arthroplasty).

The panel suggests using pneumatic compression devices over graduated compression stockings for patients undergoing major surgery, recognizing that there was very low certainty evidence for a net health benefit/harm. The recommendation applies to patients undergoing major surgery who are considered at risk for VTE.

In settings where pneumatic compression devices are not available, the use of graduated compression stockings is reasonable, because mechanical prophylaxis is an acceptable and feasible option. Further well-designed studies using clinically relevant end points are required to improve the quality of evidence related to this question. Studies outside the field of orthopedics would be particularly useful.

Question: Should combined pharmacological and mechanical prophylaxis vs pharmacological prophylaxis alone be used for patients undergoing major surgery?

We identified 7 systematic reviews addressing this question.^{23-26,28,29,31}  We identified 19 studies in these reviews that fulfilled our inclusion criteria and measured outcomes relevant to this context.^{36,60,62,68,70,104-117}  Our systematic search of RCTs did not identify any additional study that fulfilled the inclusion criteria.

Seven studies reported the effect of the combination of pharmacological and mechanical prophylaxis compared with pharmacological prophylaxis alone on the risk of mortality.^{36,60,62,68,104,105,107}  Ten studies reported the effect on the development of symptomatic PEs,^{60,62,68,70,105,107,109-111,117}  and 6 studies reported the effect on any PE.^{36,104,108,112,116,117}  Three studies reported data on symptomatic proximal DVTs,^62,70,112  and 8 studies reported data on any proximal DVT.^{62,70,104,108,112-114}  Three studies reported data on symptomatic distal DVTs,^70,105,112  and 7 studies reported on any distal DVT.^{68,105,108,109,112-114}  Six studies reported the effect of combination pharmacological and mechanical prophylaxis compared with pharmacological prophylaxis alone on the risk of major bleeding,^{60,62,68,104,109,112}  and 2 studies reported the effect on the risk of reoperation.^107,117 

The EtD framework is available online at https://guidelines.gradepro.org/profile/9AC669C6-30BB-C8DF-8430-3EDA0D4842C8.

There may be no difference in mortality between pharmacological prophylaxis combined with mechanical prophylaxis and pharmacological prophylaxis alone (RR, 0.29; 95% CI, 0.06-1.38; low certainty in the evidence of effects); this corresponds to 5 fewer (7 fewer to 3 more) deaths per 1000 patients. There may be a reduction in symptomatic PEs (RR, 0.40; 95% CI, 0.25-0.65; low certainty in the evidence of effects) favoring combined prophylaxis. Depending on the baseline risk, this benefit likely corresponds to 5 fewer (3-6 fewer) per 1000 patients with a baseline risk of 0.8% to up to 7 fewer (4-8 fewer) per 1000 patients based on a baseline risk of 1.2% from observational data.⁷³  We are very uncertain whether the risks of symptomatic proximal DVTs (RR, 0.14; 95% CI, 0.01-2.63; very low certainty in the evidence of effects) and symptomatic distal DVTs (RR, 1.99; 95% CI, 0.35-11.33; very low certainty in the evidence of effects) differ between the 2 groups.

Rates of major bleeding may be similar (RR, 1.05; 95% CI, 0.32-3.40; low certainty in the evidence of effects), corresponding to 0 fewer (5 fewer to 17 more) events per 1000 patients. We were unable to estimate the RR of major reoperation given that there were no events in either group in the 2 included trials.^107,117 

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations and imprecision.

The panel determined that there was possibly important uncertainty or variability in how much affected individuals valued the main outcomes. They further judged that the balance between desirable and undesirable effects probably favors combined pharmacological and mechanical prophylaxis over pharmacological prophylaxis alone. The panel was unable to assess the impact of adding mechanical prophylaxis on the risk of other outcomes, such as falls or skin complications. These might be considered “unmeasured harms” of mechanical prophylaxis. The panel judged the costs associated with combined prophylaxis to be moderate based on very low certainty in the evidence about resource requirements. Cost-effectiveness probably favors combined pharmacological and mechanical prophylaxis. There would probably be no impact on health equity; combined pharmacological and mechanical prophylaxis would probably be acceptable to stakeholders and probably feasible to implement. The panel determined that there was very low certainty evidence for a net health benefit/harm for combined pharmacological and mechanical prophylaxis over pharmacological prophylaxis alone. Most of the evidence evaluating this question comes from the orthopedic (joint arthroplasty) setting.

The guideline panel suggests using combined pharmacological and mechanical prophylaxis over pharmacological prophylaxis alone for patients undergoing major surgery, based on very low certainty in the evidence of effects. The panel judged that combined pharmacological and mechanical prophylaxis would be most beneficial for patients considered at very high risk for VTE following major surgery.

Further high-quality research studies using clinically important outcomes comparing combination pharmacological and mechanical methods with pharmacological methods alone are required to provide greater certainty about this recommendation. Studies addressing this question outside the orthopedic setting are most needed.

Question: Should mechanical prophylaxis combined with pharmacological prophylaxis vs mechanical prophylaxis alone be used for patients undergoing major surgery?

We identified 7 systematic reviews addressing, in part, this question.^{23-26,29,32,33}  We identified 19 studies in these reviews that fulfilled our inclusion criteria and measured outcomes relevant to this context.^{36,60,62,68,118-132}  Our systematic search of RCTs identified 5 additional studies that fulfilled the inclusion criteria.^{70,72,133-136} 

Fourteen studies reported the effect of combined mechanical and pharmacological prophylaxis compared with mechanical prophylaxis alone on the risk of mortality.^{36,60,62,68,120,121,124,125,127,128,130,133,134,136}  Sixteen studies reported the effect on the development of symptomatic PEs,^{60,62,70,120-131,134}  and 11 studies reported on the effect on any PEs^{36,118,120,121,126,128,130,131,133,135,136}  Six studies reported data on symptomatic DVTs,^{62,70,120,127,128,131}  and 10 studies reported data on any proximal DVT.^{62,70,120-122,124,125,127,128,130}  Five studies reported data on symptomatic distal DVTs,⁷⁰  and 14 studies reported on any distal DVT.^{118,120-122,124,15,127,128,130}  Fifteen studies reported the effect of combined mechanical and pharmacological prophylaxis on the risk of major bleeding,^{60,62,68,118,119,121,123-125,127,128,130,134-136}  and 4 studies reported the effect on the risk of reoperation.^{118,125,134,135} 

The EtD framework is available online at https://guidelines.gradepro.org/profile/75138F44-7AFE-A008-A8BD-10DD1DFD5377.

There is likely no difference in mortality between combined prophylaxis and mechanical prophylaxis alone (RR, 1.24; 95% CI, 0.67-2.30; moderate certainty in the evidence of effects), corresponding to 3 more (4 fewer to 17 more) deaths per 1000 patients. There is likely a reduction in symptomatic PEs (RR, 0.34; 95% CI, 0.13-0.90; moderate certainty in the evidence of effects) favoring combined prophylaxis. Depending on the baseline risk, assumed to be 0.8% or 1.1% based on a large observational study,⁷³  this benefit likely corresponds to 5 fewer (1-7 fewer) per 1000 patients in a lower-risk population to 7 fewer (1-10 fewer) per 1000 patients in a higher-risk population. The risk of symptomatic proximal DVT (RR, 0.71; 95% CI, 0.07-6.75; low certainty in the evidence of effects) and symptomatic distal DVT (RR, 0.38; 95% CI, 0.06-2.42; low certainty in the evidence of effects) may be similar between the 2 interventions, irrespective of the baseline risk group.

Combined prophylaxis likely results in a small increased risk for major bleeding (RR, 2.23; 95% CI, 1.09-4.57; moderate certainty in the evidence of effects). This likely corresponds to 14 more (1-42 more) per 1000 patients. Rates of major reoperation may be similar (RR, 2.96; 95% CI, 0.73-12.05; low certainty in the evidence of effects) between the 2 interventions, corresponding to 4 more (1 fewer to 21 more) per 1000 patients.

We rated the overall certainty in the evidence of effects as low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations and imprecision.

The panel judged that the balance between desirable and undesirable effects does not favor combined pharmacological and mechanical prophylaxis vs mechanical prophylaxis alone. Instead, the balance between desirable and undesirable effects will depend upon the risk of VTE and bleeding based on the individual patient and the type of surgical procedure. The panel judged the costs associated with combined prophylaxis to be moderate based on very low certainty in the evidence about resource requirements. The panel determined that there was possibly important uncertainty or variability in how much affected individuals valued the main outcomes. Cost-effectiveness varies based on the underlying VTE and bleeding risk categories. There would probably be no impact on health equity, and combined prophylaxis and mechanical prophylaxis alone would be acceptable to stakeholders and probably feasible to implement.

The guideline panel suggests using combined mechanical and pharmacological prophylaxis or mechanical prophylaxis alone for patients undergoing major surgery (based on low certainty in the evidence of effects). The balance of effects was considered dependent upon the risk of VTE and bleeding. Therefore, it is important to establish the baseline risk for VTE and major bleeding in surgical patients. For patients considered at high thrombosis risk and low bleeding risk, combined mechanical and pharmacological prophylaxis should be considered. For patients at high bleeding risk, mechanical prophylaxis methods alone may be preferred.

Further high-quality research studies using clinically important outcomes to identify patients with high baseline risk for VTE in whom combined pharmacological and mechanical prophylaxis would be of value, particularly outside the orthopedic setting, are needed.

Question: Should insertion of an IVC filter vs no IVC filter be used for VTE prophylaxis for patients undergoing major surgery?

We identified 2 systematic reviews of RCTs and observational studies^137,138  that addressed this research question in bariatric surgery and trauma patients. We identified 14 studies^139-152  in these reviews that fulfilled our inclusion criteria and measured outcomes relevant to this context. Our search for RCTs and observational studies identified 1 additional study that fulfilled the inclusion criteria.¹⁵³  Additionally, we identified 1 systematic review of RCTs and observational studies that was published after our initial literature search, which did not include any new study not already included in our meta-analysis.¹⁵⁴  We also identified 1 RCT published in July 2019, evaluating the use of IVC filters for patients experiencing trauma.¹⁵⁵  Although we did not update the meta-analysis, the trial results were assessed by the panel as consistent with the recommendation.

Of the 15 studies included in the meta-analysis, 1 was an RCT.¹⁵⁰  Twelve of the studies reported the effect of IVC filters on the risk of mortality.^{139,143-147,149-153}  Five studies assessed the development of symptomatic PEs,^{142,149,150,152,153}  and 10 studies assessed the development of any PE.^{139-141,143-148,151}  Ten studies assessed the development of DVTs,^{139,141,143,144,146-148,150,151,153}  with 1 study assessing the development of symptomatic DVTs specifically¹⁴¹  and 1 study assessing the development of proximal DVTs specifically.¹⁵³ 

The EtD framework is available online at https://guidelines.gradepro.org/profile/4885EDB9-B445-5554-BD62-CFE2EED6D08E.

IVC filter use may increase mortality slightly (RR, 1.38; 95% CI, 0.81-2.37; very low certainty in the evidence of effects), although the confidence interval was wide and included the possibility of no increase. Overall, we are very uncertain of this finding. Based on the control group event rate of 1.1% in this meta-analysis, this would correspond to 4 more deaths (2 fewer to 15 more) per 1000 patients receiving an IVC filter following major surgery or trauma. IVC filters may reduce the risk of symptomatic PE following major surgery and trauma (RR, 0.29; 95% CI, 0.11-0.80; very low in the evidence of effects), but we are very uncertain about this finding. Based on a baseline risk of 0.8% from observational data,⁷³  this corresponds to 6 fewer (2-7 fewer) symptomatic PEs. Based on a higher baseline risk of 1.1%,⁷³  this could result in 8 fewer PEs (2-10 fewer) per 1000 patients receiving an IVC filter; however, this is very uncertain. Rates of symptomatic proximal DVT may be increased with use of IVC filters (RR, 2.19; 95% CI, 1.07-4.50; very low certainty in the evidence of effects), but we are once again very uncertain of this finding. This corresponds to 20 more (1-58 more) or 31 more (2-92 more) per 1000 patients, based on baseline risks of 1.6% and 2.6%, respectively, from observational data.⁷³  We are also uncertain whether rates of symptomatic distal DVT are increased (RR, 2.72; 95% CI, 1.41-5.21; very low certainty in the evidence of effects), corresponding to 2 more (1-6 more) to 4 more (1-9 more) per 1000 patients, based on baseline risks of 0.1% and 0.2%, respectively, from observational data.⁷³ 

The panel did not consider potential harms of IVC filters beyond VTE. These potential harms would include potentially severe complications, such as IVC perforation and IVC filter embolization.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, inconsistency, and indirectness.

The panel determined that there was possibly important uncertainty or variability in how much affected individuals might value the main outcomes. They judged that the balance between desirable and undesirable effects favored not using IVC filters in the setting of major surgery or trauma. The panel judged from our analysis of data, largely from observational studies, that the high rates of DVT and the trend for higher mortality associated with the use of IVC filters outweighed the potential reduction in PEs. Furthermore, a recently published high-quality RCT of IVC filters, following major trauma for patients in whom pharmacological prophylaxis was considered contraindicated, did not find that IVC filters reduced symptomatic PE or death.¹⁵⁵  Given there are serious nonthrombotic risks associated with IVC filters (eg, IVC perforation and IVC filter embolization) that were not considered in our analysis, this would further strengthen our recommendation against IVC filter use.¹⁵⁶ 

The panel also recognized that the cost of IVC filters and resources associated with their insertion were large. The panel concluded that routine use of IVC filters following major surgery or trauma was probably not cost-effective and favored not placing a filter. Because of the increased resources required, high costs, and limited availability of interventional radiology units, health equity would probably be reduced by use of IVC filters. Routine use of IVC filters might also not be acceptable or feasible to implement for some stakeholders.

The panel suggests against using IVC filters for prophylaxis of VTE for patients undergoing major surgery or trauma patients based upon very low certainty in the evidence. The very low quality evidence underlying this recommendation supports the need for well-designed RCTs evaluating clinically important outcomes to better define the role of IVC filters for patients who cannot safely or feasibly receive pharmacological or mechanical prophylaxis following major surgery or trauma. Further studies quantifying the nonthrombotic risks of IVC filters would also be of value.

Question: Should extended antithrombotic prophylaxis vs short-term antithrombotic prophylaxis be used for patients undergoing major surgery?

We identified 9 systematic reviews addressing this research question.^157-165  We identified 14 studies^166-179  in these reviews that fulfilled our inclusion criteria and measured outcomes relevant to this context. Our systematic search of RCTs identified 6 additional studies¹⁸⁰  not included in previous systematic reviews that fulfilled the inclusion criteria.

Seventeen studies reported the effect of extended vs short-term duration of pharmacological thromboprophylaxis on the development of mortality,^{166,167,170-173,175-185}  17 studies reported the effect on the development of PEs,^{166,167,170-184}  18 studies reported the effect on the development of proximal DVTs,^{166-173,175-184}  14 studies reported the effect on the development of distal DVTs,^{166,169,171-173,175,177-184}  16 studies reported the effect on the risk of major bleeding,^{167-173,177-185}  and 6 studies reported the effect on the risk of reoperation.^{166,173,174,179,184,185}  In general, these studies compared shorter courses of pharmacological prophylaxis (4-14 days) with extended courses of pharmacological prophylaxis (19-42 days) and then followed patients for a common period (3-9 months) for VTE and bleeding complications.

The EtD framework is available online at https://guidelines.gradepro.org/profile/79bce70d-c689-4fbf-b0e4-c2ec3142bb2c.

There is likely no difference in mortality between extended- and standard-course antithrombotic prophylaxis (RR, 0.94; 95% CI, 0.64-1.39; moderate certainty in the evidence of effects); this corresponds to 1 fewer death (6 fewer to 6 more) per 1000 patients. There is likely a small reduction in symptomatic PEs (RR, 0.44; 95% CI, 0.22-0.85; moderate certainty in the evidence of effects). Depending on baseline risk,⁷³  this corresponds to 4 fewer (1-6 fewer) deaths per 1000 patients with a baseline risk of 0.8% and 6 fewer (2-9 fewer) deaths per 1000 patients with a baseline risk of 1.1% receiving extended pharmacological prophylaxis. Symptomatic proximal DVTs are also likely reduced (RR, 0.30; 95% CI, 0.21-0.42; moderate certainty in the evidence of effects). Depending upon the baseline risk,⁷³  this corresponds to 12 fewer (10-13 fewer) symptomatic proximal DVTs per 1000 patients in a lower-risk group of patients with a 1.6% baseline risk or 18 fewer (15-21 fewer) per 1000 patients in a higher-risk group with a 2.6% baseline risk. Extended pharmacological prophylaxis likely reduces distal DVTs (RR, 0.57; 95% CI, 0.37-0.87; moderate certainty in the evidence of effects), which corresponds to 1 fewer (0-1 fewer) symptomatic distal DVT per 1000 patients based on a 0.1% baseline risk from observational data.⁷³ 

Rates of major bleeding may be similar (RR, 1.00; 95% CI, 0.59-1.70; low certainty in the evidence of effects), corresponding to 0 fewer (3 fewer to 6 more) per 1000 patients. Rates of reoperation may also be similar (RR, 0.82; 95% CI, 0.34-1.99; very low certainty in the evidence of effects), but we are very uncertain about this finding.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations and very serious imprecision.

The panel determined that there was possibly important uncertainty or variability in how much affected individuals valued the main outcomes. They further judged that the balance between desirable and undesirable effects probably favors extended-duration vs standard-duration prophylaxis. The panel judged the costs associated with extended-duration prophylaxis to be moderate based on very low certainty in the evidence. Cost-effectiveness probably favors extended-duration prophylaxis. Health equity is possibly reduced with extended-duration prophylaxis, with economically disadvantaged patients potentially being unable to afford the required medications or medication copayments. Extended-duration prophylaxis would probably be acceptable to stakeholders and probably feasible to implement. The panel recognized that most of the trials compared a prolonged prophylaxis (up to 42 days, or ∼6 weeks following surgery) with a short duration of anticoagulant prophylaxis designed to approximate the length of a postoperative hospital stay (∼4-14 days) in the eras in which the studies were performed. Furthermore, the panel recognized that these studies were largely limited to 2 high-risk surgical scenarios (total hip or knee arthroplasty and major cancer general surgical procedures).

The guideline panel determined that the net benefit favored using extended-course antithrombotic prophylaxis over short-term antithrombotic prophylaxis for patients undergoing major surgery based on very low certainty evidence. Given the very low certainty in the evidence of effects this is based upon, there is a critical need for higher-quality studies comparing extended vs short-term prophylaxis using clinically important outcomes in contemporary surgical practices, which are marked by early patient mobilization and shorter hospital stays. There is particularly a need for studies outside the general hip and knee arthroplasty and cancer general surgical settings to confirm the benefits of extended prophylaxis in other settings. There also appears to be a need for further research to determine the optimal duration of extended prophylaxis.

Question: Should early vs delayed antithrombotic prophylaxis be used for patients undergoing major surgery?

We did not find any systematic reviews that addressed this question. In our systematic search of the literature we found 6 studies that fulfilled our inclusion criteria and measured outcomes relevant to this context.^186-191  We cross-referenced the studies found in our search with the references from a recent narrative review¹⁹²  but did not identify any additional studies that fulfilled our inclusion criteria. All studies included surgical patients. Twelve hours was selected to be the cutoff point between early and late postsurgical antithrombotic administration. Six studies^186-191  reported the effect of early vs late postsurgical antithrombotic administration on the risk of mortality and on the risk of development of any PEs. Five studies^186,188-191  reported the effect on the risk of any proximal and any distal DVTs, and 1 reported the effect on the risk of any DVT.¹⁸⁷  Only the 5 studies that specified the location of the DVT were included in the evidence profile. Six studies^186-191  reported the effect of early vs late postsurgical antithrombotic administration on the risk of major bleeding and on the risk of reoperation.

The EtD framework is available online at https://guidelines.gradepro.org/profile/E664E38D-FA7C-DBC9-8E77-373D0582050E.

Early prophylaxis may result in no difference in mortality (RR, 1.57; 95% CI, 0.77-3.19; very low certainty in the evidence of effects), but we are very uncertain about this finding. This corresponds to 1 more (1 fewer to 6 more) per 1000 patients. We are uncertain about the effect of early prophylaxis on symptomatic PEs (RR, 0.63; 95% CI, 0.23-1.72; very low certainty in the evidence of effects); depending on the baseline risk from observational data,⁷³  this corresponds to 3 fewer (6 fewer to 6 more) to 4 fewer (8 fewer to 8 more) per 1000 patients with baseline risks of 0.8% and 1.1%, respectively. We are also uncertain of the effect of early prophylaxis on symptomatic proximal DVTs (RR, 0.88; 95% CI, 0.40-1.96; very low certainty in the evidence of effects), corresponding to 2 fewer (10 fewer to 16 more) to 3 fewer (16 fewer to 25 more) per 1000 patients when applying baseline risks of 1.6% and 2.6%, respectively.⁷³  Early prophylaxis has an uncertain effect on distal DVTs (RR, 0.68; 95% CI, 0.41-1.12; very low certainty in the evidence of effects), with an absolute risk reduction from 0 fewer (0-1 fewer; baseline risk, 0.1%⁷³ ) to 1 fewer (0-1 fewer; baseline risk, 0.2%⁷³ ) symptomatic distal DVT per 1000 patients.

The risk of major bleeding may be similar (RR, 1.63; 95% CI, 0.81-3.29; very low certainty in the evidence of effects), corresponding to 5 fewer (1 fewer to 17 more), although we are very uncertain of this finding. Also, the need for reoperation may be increased (RR, 1.84; 95% CI, 0.89-3.80; very low certainty in the evidence of effects) corresponding to 2 more reoperations (0 fewer to 6 more) per 1000 patients. Once again, we are very uncertain about this finding.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes and downgrading for study limitations, indirectness, and imprecision.

Based on very low certainty in the evidence of effects, the panel judged that the balance of effects did not favor early or delayed institution of pharmacological antithrombotic prophylaxis in major surgical patients. Costs and saving would likely be negligible, and the panel saw no particular issues related to equity, acceptability, or feasibility.

The guideline panel suggests early administration (postoperative, within 12 hours) or late administration (postoperative, after 12 hours) of antithrombotic prophylaxis in major surgical patients, based on very low certainty in the evidence of effects. In light of the very low certainty in the evidence of effects, further high-quality studies using clinically important outcomes are important to provide greater certainty about the benefits and risks of early pharmacological prophylaxis. The panel was particularly interested in seeing future high-quality studies of early vs late pharmacological prophylaxis studies in high-risk bleeding patients, examining the benefits and risks of later intervention (days following surgery) once the bleeding risk had greatly subsided.

Question: Should ASA vs anticoagulants be used for patients undergoing total hip or knee arthroplasty?

We found 7 RCTs that compared the use of ASA vs anticoagulants for patients undergoing total hip arthroplasty or total knee arthroplasty.^193-199  Additionally, we identified 2 trials comparing ASA with LMWH in total hip arthroplasty patients²⁰⁰  and ASA with DOAC in total hip arthroplasty or total knee arthroplasty patients,²⁰¹  in which all participants received a 10-day period of LMWH or a 5-day period of DOACs, respectively, prior to randomization. The trials were reviewed by the panel but were not included in the main meta-analysis because of differences in the comparator groups. Of the 7 studies included in the analysis, 2 studies compared ASA with UFH,^193,195  4 studies compared ASA with LMWH,^{194,196,198,199}  and 2 studies compared ASA with oral anticoagulants.^197,199  All 7 studies reported the outcomes of mortality and PE,^193-199  6 studies reported on proximal and distal DVTs,^195-199  and 5 studies reported on major bleeding.^{194-196,198,199}  We found no studies addressing the outcome of reoperation.

The EtD framework is available online at https://guidelines.gradepro.org/profile/3532ED1D-6A40-A982-BC3F-6DA318B3B611.

There may be no difference in mortality between ASA and anticoagulants (RR, 2.32; 95% CI, 0.15-36.90; low certainty in the evidence of effects). There also may be no difference in the risk of symptomatic PEs between ASA and anticoagulants (RR, 1.49; 95% CI, 0.37-6.09; very low certainty in the evidence of effects), corresponding to 3 more (4 fewer to 29 more) symptomatic PEs per 1000 patients based on a baseline risks of 0.6%^202,203  from observational data; however, we are very uncertain of this finding. There also may be no difference in the risk of proximal DVTs (RR, 1.49; 95% CI, 0.51-4.34; very low certainty in the evidence of effects), corresponding to 3 more symptomatic (3 fewer to 30 more) proximal DVTs per 1000 patients based on a baseline risk of 0.6%,^202,203  or in the risk of distal DVTs (RR, 1.45; 95% CI, 0.86-2.46; very low certainty in the evidence of effects), corresponding to 0 fewer symptomatic distal DVTs (0 fewer to 1 more) per 1000 patients based on a baseline risk of 0.05%,^202,203  although we were very uncertain about both findings.

ASA may lead to a small increased risk for major bleeding (RR, 2.63; 95% CI, 0.64-10.79; low certainty in the evidence of effects). These findings correspond to 6 more (1 fewer to 35 more) major bleeding events per 1000 patients. We found no evidence to inform the comparative risk of reoperation.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, indirectness, and imprecision.

The panel determined that there was probably important uncertainty or variability in how much affected individuals value the main outcomes. It further judged that use of ASA saved costs and resources; however, the results of cost-effectiveness studies varied, with some favoring ASA and others favoring anticoagulant prophylaxis. Health equity would probably be increased as a consequence of the use of ASA. Acceptability was thought to also vary depending on the type of stakeholder (patient vs health care provider). There were no concerns about the feasibility of implementation. A recent large RCT supports our recommendation that ASA or anticoagulants be used for VTE prophylaxis following total hip or knee arthroplasty. However, this study was not included in our analysis because all patients received a 5-day course of a DOAC before being randomized to ASA or to stay on a DOAC for extended prophylaxis.²⁰¹ 

The guideline panel suggests using ASA or anticoagulants for patients undergoing total hip arthroplasty or total knee arthroplasty (conditional recommendation based on very low certainty in the evidence of effects). They determined that there was very low certainty evidence for any net health benefit/harm from using ASA vs anticoagulants. Of 8 panel members who voted on this recommendation, 5 voted for recommending either intervention, and 3 voted for a conditional recommendation in favor of anticoagulants.

The panel identified that there is a need for large well-designed clinical trials using clinically important end points comparing ASA with other pharmacological methods following total hip and knee arthroplasty. The panel noted that such studies are underway.

Question: Should DOACs vs LMWH be used for patients undergoing total hip or knee arthroplasty?

We identified 1 systematic review²⁰⁴  that addressed this question. Twenty-two studies in this review fulfilled our inclusion criteria. Our update of the systematic review identified 16 additional studies. All studies included patients undergoing elective hip or knee replacement.

Five studies assessed the effects of dabigatran,^191,205-208  15 studies assessed the effects of rivaroxaban,^{180,185,189,190,199,209-218}  4 studies assessed the effects of apixaban,^219-222  5 studies assessed the effects of darexaban^223-226  and edoxaban,^227-231  and 4 studies assessed the effects of other DOACs.^232-235 

Thirty-four studies reported mortality,^{180,185,189-191,199,205-215,218-222,224-231,233-235}  whereas 33 studies reported nonfatal PEs.^{180,185,189-191,199,205-212,214-216,219-222,224-231,233-235}  We estimated proximal and distal DVTs using the pooled estimate from symptomatic DVTs, which was reported in 30 studies.^{185,189-191,199,205-212,214-216,219-222,224-230,234,235}  Thirty-two studies reported major bleeding,^{180,185,189-191,205-212,214,215,219-222,224-235}  whereas only 15 studies reported bleeding leading to reoperation.^{180,185,189-191,205-207,209-212,215,220,221} 

We tested potential differences in the effects with specific drugs and between classes (anti–factor IIa vs anti–factor Xa). We found no interaction for any of the outcomes. Additionally, we conducted a sensitivity analysis excluding dose-finding studies. The results did not change appreciably.

The EtD framework is available online at https://guidelines.gradepro.org/profile/9160FAA2-4F98-A3AA-9816-64DF796ABBC7.

DOACs probably do not reduce mortality compared with that associated with LMWH (RR, 0.94; 95% CI, 0.53-1.66; moderate certainty in the evidence of effects); this corresponds to 0 fewer deaths (1 fewer to 1 more) per 1000 patients. DOACs probably slightly reduce the rate of symptomatic PEs (RR, 0.74; 95% CI, 0.50-1.10; moderate certainty in the evidence of effects); based on a baseline risk of 0.6% from observational data,^202,203  this corresponds to 1 fewer (3 fewer to 1 more) symptomatic PE per 1000 patients. The use of DOACs reduces symptomatic proximal DVTs slightly (RR, 0.56; 95% CI, 0.39-0.79; high certainty in the evidence of effects), which corresponds to 3 fewer (1-4 fewer) symptomatic proximal DVTs per 1000 patients, based on a baseline risk of 0.6% from observational data.^202,203  This effect on symptomatic distal DVTs is probably not clinically relevant (RR, 0.56; 95% CI, 0.39-0.79; high certainty in the evidence of effects), which corresponds to 0 fewer symptomatic distal DVTs per 1000 patients based on a baseline risk of 0.049%, from observational data.^202,203 

DOACs probably do not increase major bleeding compared with LMWH (RR, 1.03; 95% CI, 0.79-1.35; moderate certainty in the evidence of effects), which corresponds to 0 fewer major bleeding events (2 fewer to 4 more) per 1000 patients. Similarly, rates of reoperation may not be meaningfully increased (RR, 1.43; 95% CI, 0.75-2.71; moderate certainty in the evidence of effects) given the low event rates; this corresponds to 0 fewer reoperations (0 fewer to 2 more) per 1000 patients.

We rated the overall certainty in the evidence of effects as moderate based on the lowest certainty in the evidence for the critical outcomes, downgrading for imprecision.

The panel judged the desirable and undesirable effects as being small and trivial, respectively, in magnitude. Cost-effectiveness was judged to probably favor the use of DOACs. Similarly, equity, acceptability, and feasibility each favored the use of DOACs and contributed to the recommendation in their favor. Use of out-of-hospital prophylaxis, which is routine following total hip or knee arthroplasty, particularly favored DOACs over LMWH, given the need for parenteral administration of the latter agent.

The guideline panel suggests using DOACs rather than LMWH for patients undergoing total hip or knee arthroplasty. Based on an overall moderate certainty in the evidence of effects, the panel judged the balance of effects to probably favor the use of DOACs over LMWH. The ultimate judgment of a conditional recommendation for DOACs was based on anonymous voting by panel members without direct financial conflicts, with a majority of 5 voting for this recommendation (vs 4 in favor of a recommendation for using either). The panel recommended a need for large clinical trials using clinically relevant end points comparing different DOACs. Further studies regarding the optimal timing of the initiation of postoperative dosing of DOACs are warranted.

Question: Should 1 DOAC vs another DOAC be used for patients undergoing total hip or knee arthroplasty?

We found no study that compared different classes of DOACs or individual DOACs of the same class head to head. Therefore, we used the body of evidence comparing DOAC prophylaxis vs LMWH (see recommendation 10) as the basis for an indirect assessment of their relative effectiveness.

The EtD framework is available online at https://guidelines.gradepro.org/profile/16A3927F-FB2D-C278-9FE5-D4535020FE27.

In the absence of comparative trials of different classes of DOACs (anti–factor IIa vs anti–factor Xa) or individual DOACs of the same class, we tested potential differences by analyzing potential subgroup effects. Based on the tests for interaction, we did not demonstrate any evidence for a clinically relevant subgroup effect for any of the potentially desirable outcomes. Based on this finding, the panel assumed that the beneficial effects were likely similar for different DOACs.

Similarly, based on tests for interaction, we did not find any evidence for a clinically relevant subgroup effect for any of the potentially undesirable outcomes. Based on this finding, the panel assumed that the undesirable effects were likely similar for different DOACs.

We rated the overall certainty in the evidence of effects as low based on the lowest certainty in the evidence for the critical outcomes, downgrading for indirectness and imprecision.

Based on these findings, the panel judged that the balance of effects did not favor any particular DOAC over another. The panel also judged that issues surrounding cost-effectiveness, equity, acceptability, and feasibility also did not weigh in sufficiently to favor 1 DOAC over another.

Based on overall low certainty in the evidence of effects, the panel judged that there were no net benefits in favor of any DOAC vs another. Given the lack of direct comparative evidence, the panel identified an important need for high-quality head-to-head studies comparing different DOACs for the prevention of VTEs following total hip or knee arthroplasty.

Question: Should LMWH vs warfarin be used for patients undergoing total hip or knee arthroplasty?

We identified 1 systematic review that addressed this question.²³⁶  We identified 7 trials in this review that fulfilled our inclusion criteria and measured outcomes relevant to this context.^{39,219,237-241}  Our systematic search of RCTs did not identify any additional study that fulfilled the inclusion criteria. Three studies were conducted with patients undergoing total hip arthroplasty,^39,239,241  3 studies were conducted with patients undergoing total knee arthroplasty,^219,237,240  and 1 study addressed both populations.²³⁸  Five studies^219,237-240  reported the effect of LMWH compared with warfarin on mortality, 5 studies reported the effect on the development of symptomatic PEs,^{39,219,237,238,240}  6 studies reported on any proximal DVT,^219,237-241  and 2 studies reported on any distal DVT.^237,241  All 7 studies reported the effect on the risk of major bleeding,^{39,219,237-241}  and 2 studies reported on the risk of reoperation.^237,241 

The EtD framework is available online at https://guidelines.gradepro.org/profile/BC1783C1-D62B-AECB-B9F7-87A9D474A834.

LMWH may result in little or no difference in mortality compared with warfarin (RR, 0.51; 95% CI, 0.14-1.88; low certainty in the evidence of effects). LMWH likely does not reduce symptomatic PEs (RR, 0.83; 0.27-2.54; moderate certainty in the evidence of effects). LMWH may reduce symptomatic proximal DVTs (RR, 0.61; 95% CI, 0.36-1.02; very low certainty in the evidence of effects) and symptomatic distal DVTs (RR, 0.61; 95% CI, 0.42-0.88; low certainty in the evidence of effects). This corresponds to 2 fewer (0-4 fewer) symptomatic proximal DVTs and 0 fewer symptomatic distal DVTs with the use of LMWH than with warfarin for 1000 patients treated, based on baseline risks of 0.6% and 0.049%, respectively, from observational data.^202,203 

LMWH use likely results in increased major bleeding compared with the use of warfarin (RR, 1.81; 95% CI, 1.31-2.50; moderate certainty in the evidence of effects). This corresponded to 16 more (5-22 more) major bleeds per 1000 patients. There was no difference with regard to reoperation rates between those receiving LMWH or warfarin (RR, 3.09; 95% CI, 0.13-75.48; moderate certainty in the evidence of effects), which corresponded to 0 more events per 1000 patients.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations and very serious imprecision.

The panel judged that there was possibly important uncertainty or variability in how much people value the main outcomes. The balance between desirable und undesirable effects probably favored LMWH. Resources requirements of warfarin were deemed moderate, particularly with regard to the need for, and the complexity of, anticoagulant monitoring, but cost-effectiveness data probably did not favor warfarin or LMWH. There probably would be no impact on health equity; both agents appear acceptable to stakeholders and are feasible to implement.

The guideline panel suggests using LMWH rather than warfarin for patients undergoing total hip arthroplasty or total knee arthroplasty. The guideline panel determined that there was very low certainty evidence for a net health benefit/harm from using LMWH rather than warfarin. Based on the body of available evidence, it is likely that warfarin reduces the risk of major bleeding based on evidence of moderate certainty. However, it may also increase the risk of proximal DVTs, based on very low quality evidence.

Further high-quality studies using clinically important outcomes would be of value to improve the certainty in the recommendation. However, given the availability of DOACs as oral agents that do not require anticoagulant monitoring or dose adjustment, further clinical trials using warfarin are not regarded as a high priority at this time.

Question: Should LMWH vs UFH be used for patients undergoing total hip or knee arthroplasty?

We identified 1 systematic review that addressed this question.²³⁶  We identified 12 trials in this review that fulfilled our inclusion criteria and measured outcomes relevant to this context.^242-253  Our systematic search of RCTs did not identify any additional study that fulfilled the inclusion criteria. Ten trials were performed on patients undergoing total hip arthroplasty,^{242-244,246,247,249-253}  and 2 trials were conducted on patients undergoing total knee arthroplasty.^245,248  Five trials reported the effect of LMWH compared with UFH on mortality,^{244,247,249,251,252}  10 studies reported the effect on the development of symptomatic PEs,^{242-245,247-249,251-253}  8 studies reported the effect on any proximal DVT,^{242,244-249,251}  and 6 studies reported the effect on any distal DVT.^{242,244-249,251}  Six studies reported the effect on the risk of major bleeding,^{244,245,249-251,253}  and 2 studies reported the effect on the risk of reoperation.^247,248 

The EtD framework is available online at https://guidelines.gradepro.org/profile/06FDBFB0-4D4E-E0D0-AEAD-C8B371DFA939.

LMWH results in little or no difference in mortality compared with UFH (RR, 0.26; 95% CI, 0.03-2.36; high certainty in the evidence of effects), which corresponded to 3 fewer (4 fewer to 5 more) deaths per 1000 patients. LMWH probably reduces the risk of symptomatic PEs slightly (RR, 0.37; 95% CI, 0.19-0.71; moderate certainty in the evidence of effects). This corresponds to 4 fewer (2-5 fewer) symptomatic PEs per 1000 patients, based on a baseline risk of 0.6% from observational data.^202,203  LMWH also likely reduces the risk of symptomatic proximal DVTs (RR, 0.48; 95% CI, 0.34-0.69; moderate certainty in the evidence of effects), corresponding to 3 fewer (2-4 fewer) per 1000 patients, based on a baseline risk of 0.6% from observational data.^202,203  LMWH appears to result in little or no difference in symptomatic distal DVTs (RR, 1.18; 95% CI, 0.81-1.72; low certainty in the evidence of effects), with very small corresponding absolute effect size estimates based on a baseline risk of 0.049% from observational data.^202,203 

LMWH likely results in a small decrease in the risk of major bleeding (RR, 0.55; 95% CI, 0.27-1.13; moderate certainty in the evidence of effects); this corresponds to 19 fewer (30 fewer to 5 more) major bleeds per 1000 patients. We were unable to estimate an effect on the risk of reoperation given that the included studies reported no events for this outcome.

We rated the overall certainty in the evidence of effects as moderate based on the lowest certainty in the evidence for the critical outcomes, downgrading for imprecision.

The panel determined that there was probably important uncertainty or variability in how much affected individuals value the main outcomes. Cost-effectiveness likely differs by country but probably favors LMWH. The panel assessed that this recommendation probably would have no impact on health equity and would be acceptable to stakeholders. LMWH is already widely used, and the panel had no concern about the feasibility of implementation.

The guideline panel recommends LMWH rather than UFH for patients undergoing total hip arthroplasty or total knee arthroplasty.

The guideline panel determined that there is moderate certainty evidence for a net health benefit/harm from using LMWH over UFH. Future large studies using clinically relevant end points would help to better inform this recommendation, although this research question would not be regarded as high priority.

Question: Should pharmacological prophylaxis vs no pharmacological prophylaxis be used for patients undergoing hip fracture repair?

We identified 1 systematic review²⁵⁴  that addressed, in part, this question. We identified 5 studies in this review that fulfilled our inclusion criteria and measured outcomes relevant to this context.^255-259  Our systematic search of RCTs identified 7 additional studies that fulfilled the inclusion criteria.^260-266  Nine studies reported the effect of pharmacological prophylaxis compared with no intervention on risk of mortality.^{255,256,258,259,261-265}  Nine studies reported the effect on development of any PEs,^{255-259,261,264-266}  and 5 studies reported the effect of any proximal DVT and any distal DVT.^{257,260,261,263,265}  Eleven studies reported the effect of pharmacological prophylaxis compared with no intervention on risk of major bleeding,^{255,256,258-266}  and 3 studies reported the effect on risk of reoperation.^258,261,266 

We tested potential differences in the effects with ASA and anticoagulant prophylaxis and performed a subgroup analysis. The analysis indicated no subgroup effect with regard to desirable and undesirable effects comparing ASA with anticoagulant prophylaxis. As a result, in this analysis, studies with ASA are pooled with those of anticoagulant prophylaxis compared with no pharmacological prophylaxis.

The EtD framework is available online at https://guidelines.gradepro.org/profile/A10CDC06-B411-D572-959A-A8405E1373A1.

Pharmacological prophylaxis appears to result in little or no difference in mortality (RR, 0.95; 95% CI, 0.84-1.07; very low certainty in the evidence of effects), although we are very uncertain about this finding. This would correspond to 4 fewer deaths (11 fewer to 5 more) per 1000 patients based on a baseline risk of 7.1% from the control group event rate in the meta-analysis. Pharmacological prophylaxis may reduce symptomatic PEs (RR, 0.49; 95% CI, 0.33-0.72; very low certainty in the evidence of effects), but we are very uncertain of this finding. This corresponds to 6 fewer (3-7 fewer) events per 1000 patients based on a baseline risk of 1.1% from the control group event rate in the meta-analysis. Based on lower baseline risk of 0.3% from observational data,²⁶⁷  this would correspond to 2 fewer (1-2 fewer) symptomatic PEs per 1000 patients. Pharmacological prophylaxis may reduce symptomatic proximal DVTs (RR, 0.51; 95% CI, 0.38-0.69; very low certainty in the evidence of effects), but we are very uncertain of this finding. In a moderate-risk population with a baseline risk of 2.5%,²⁶⁷  this corresponds to 12 fewer (8-16 fewer) per 1000 patients. Pharmacological prophylaxis likely has little or no effect on symptomatic distal DVTs (RR, 0.85; 95% CI, 0.56-1.29; very low certainty in the evidence of effects), but once again we are very uncertain of this finding.

Pharmacological prophylaxis may increase major bleeding (RR, 1.24; 95% CI, 1.12-1.37; low certainty in the evidence of effects). Depending on baseline risk, this corresponds to 1 more (1-2 more) major bleed per 1000 patients in a lower-risk population (baseline risk of 0.5% from observational data)²⁶⁷  or as many as 20 more (10-31 more) per 1000 patients in a higher-risk population (baseline risk of 8% from the control group event rate in the meta-analysis). Pharmacological prophylaxis appears to have little or no effect on the need for reoperation (RR, 1.05; 95% CI, 0.82-1.35; very low certainty in the evidence of effects); however, we are very uncertain of this finding.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, indirectness, and very serious imprecision.

The panel judged the magnitude of the desirable effects as moderate in size and the undesirable effects as small in size. There was possibly important uncertainty or variability about how patients may value these outcomes. There would probably be no impact on equity, and the panel foresaw no issues with regard to acceptability and feasibility of using pharmacological prophylaxis in this patient population.

Although the overall certainty in the evidence of effects was very low, the panel judged that the balance of effects probably favored the use of pharmacological prophylaxis for VTE prophylaxis following hip fracture repair. Given the overall very low certainty in the evidence, the panel indicated that there remains an important need for large high-quality RCTs using clinically important end points to determine the optimal role of ASA or anticoagulant pharmacological prophylaxis in this patient population. However, higher priority would be comparative studies of different antithrombotic regimens for the prevention of VTEs in these patients requiring repair of hip fracture.

Question: Should LMWH prophylaxis vs UFH prophylaxis be used for patients undergoing hip fracture repair?

We identified 1 systematic review²⁵⁴  that addressed this question. We identified 3 studies in this review that fulfilled our inclusion criteria and measured outcomes relevant to this context.^268-270  Our update of the systematic review did not identify any additional study that fulfilled the inclusion criteria. Two studies reported the effect of LMWH prophylaxis compared with UFH prophylaxis on risk of mortality, as well as any proximal and distal DVTs,^268,269  and 3 studies reported the effect on any PEs and major bleeding.^268-270  No information on reoperation rates was available in any of the included studies.

The EtD framework is available online at https://guidelines.gradepro.org/profile/80C377E5-E3C0-36CD-B646-C2532AB4D4B9.

LMWH appears to result in little or no difference in mortality compared with UFH prophylaxis following hip fracture repair (RR, 0.47; 95% CI, 0.10-2.12; very low certainty in the evidence of effects), although we are very uncertain about the effect. Assuming a baseline risk of 7.4% for UFH-treated patients, this would correspond to 39 fewer (66 fewer to 88 more) deaths per 1000 patients. We are very uncertain about the effect of LMWH on symptomatic PEs (RR, 2.13; 95% CI, 0.06-81.3; very low certainty in the evidence of effect). LMWH may result in a small, possibly unimportant, increase in symptomatic proximal DVTs (RR, 2.24; 95% CI, 0.92-5.43; low certainty in the evidence of effect) corresponding to 31 more (2 fewer to 111 more) per 1000 patients based on a baseline risk of 2.5% from observational data.²⁶⁷  LMWH appears to result in little or no difference in symptomatic distal DVTs (RR, 0.66; 95% CI, 0.21-2.17; very low certainty in the evidence of effects).

LMWH appears to result in little or no difference in major bleeding compared with UFH after hip fracture surgery (RR, 0.85; 95% CI, 0.19-3.79; very low certainty in the evidence of effects). This corresponds to 9 fewer (50 fewer to 173 more) major bleeds per 1000 moderate-risk patients with a baseline risk of 6.2% based on a lower baseline risk of 0.5% from observational data²⁶⁷ ; the corresponding absolute risk reduction would be 1 fewer (4 fewer to 14 more) per 1000 patients. No comparative information is available regarding the risks of reoperation following hip fracture with the use of LMWH or UFH.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, indirectness, and very serious imprecision.

The panel determined that there was possibly important uncertainty or variability in how much affected individuals valued the main outcomes. They further judged that the balance between desirable and undesirable effects did not favor LMWH or UFH following hip fracture repair. The panel recognized the very low certainty in comparative evidence, which was based on three small RCTs that did not report symptomatic DVT outcomes. The panel recognized that the comparative resources associated with LMWH and UFH prophylaxis were probably negligible. Cost-effectiveness was considered to favor LMWH based upon results of a single study.²⁷¹  Using LMWH or UFH would probably not impact health equity, and either drug was deemed to be probably acceptable to stakeholders and feasible to implement.

Taking into consideration the very low certainty in the evidence, the panel judged that LMWH or UFH prophylaxis could be recommended following hip fracture repair. Large RCTs using clinically important outcomes are needed to better define the relative benefits and risks of LMWH compared with UFH following hip fracture surgery.

Question: Should pharmacological prophylaxis vs no pharmacological prophylaxis be used for patients undergoing major general surgery?

We identified a systematic review of RCTs²⁷²  addressing this research question. We identified 3 studies^273-275  in this review that fulfilled our inclusion criteria and measured outcomes relevant to this context. Our systematic search for RCTs identified 32^{49,122,125,131,274,276-302}  additional studies that fulfilled the inclusion criteria, including patients undergoing major general surgery. Sixteen studies^{122,125,274-278,284,285,291,293,296,297,299,300,302}  reported the effect of pharmacological prophylaxis vs no pharmacological prophylaxis on the risk of mortality, and 11 studies^{125,273,276-279,285,286,289,293,295}  reported the effect on the development of symptomatic PEs. Six studies^{49,125,278,285,287,290}  reported the effect on development of screening-detected proximal DVTs, and 6 studies^{124,277,283,284,286,289}  reported the effect on development of screening-detected distal DVTs.

Twelve studies^{125,273-276,282,285,287,293,295,297,298}  reported the effect of pharmacological prophylaxis vs no pharmacological prophylaxis on risk of major bleeding, and 3 studies^275,282,295  reported the effect on risk of reoperation.

The EtD framework is available online at https://guidelines.gradepro.org/profile/3B5A5678-B1D9-4D60-8E1F-F3AD700132F8.

Pharmacological prophylaxis compared with no pharmacological prophylaxis probably reduces mortality (RR, 0.76; 95% CI, 0.61-0.93; moderate certainty in the evidence of effects). This finding corresponds to 4 fewer deaths (1-7 fewer) per 1000 patients undergoing major general surgery. Pharmacological prophylaxis probably also reduces symptomatic PEs (RR, 0.45; 95% CI, 0.23-0.88; moderate certainty in the evidence of effects). This corresponds to 4 fewer (1-6 fewer) pulmonary embolic events per 1000 patients undergoing major general surgery. Pharmacological prophylaxis may also reduce symptomatic proximal DVTs (RR, 0.38; 95% CI, 0.14-1.00; very low certainty in the evidence of effects), but we are very uncertain of this finding. This corresponds to 10 fewer (0-14 fewer) symptomatic proximal DVTs per 1000 patients based on a baseline risk of 1.6% from observational data.⁷³  It may reduce symptomatic distal DVTs (RR, 0.57; 95% CI, 0.36-0.90; low certainty in the evidence of effects), which corresponds to 1 fewer (0-1 fewer) symptomatic distal DVT per 1000 patients undergoing major general surgery based on a baseline risk of 1.6% from observational data.⁷³ 

Pharmacological prophylaxis probably increases major bleeding (RR, 1.37; 95% CI, 0.89-2.13; moderate certainty in the evidence of effects). This corresponds to 10 more (3 fewer to 29 more) major bleeding events per 1000 patients undergoing major general surgery. Pharmacological prophylaxis results in little or no difference in reoperation (RR, 0.75; 95% CI, 0.21-2.77; low certainty in the evidence of effects).

The overall certainty of the estimates of effects was based on the low certainty outcomes and was not based on the lowest certainty of evidence for the critical outcomes. In this case, the recommendation was sufficiently supported by the favorable impact on desirable effects for which there was higher quality evidence.

The panel judged the desirable effects to be of moderate magnitude and the undesirable effects to be of small magnitude. They assumed that there was possibly important uncertainty or variability in patients’ values. Pharmacological prophylaxis probably would incur moderate additional costs but was judged to be probably cost-effective. Pharmacological prophylaxis would probably have no impact on equity, was probably acceptable, and was likely feasible.

The panel judged that the overall balance of effects favored pharmacological prophylaxis over no pharmacological prophylaxis for patients undergoing major general surgery based on low certainty in the evidence of effects. Further high-quality comparative studies, using appropriate clinical outcomes, would be of value to add more certainty to these recommendations. However, such studies would not be considered as high priority by the panel.

Question: Should LMWH prophylaxis vs UFH prophylaxis be used for patients undergoing major general surgery?

We identified 2 systematic reviews of RCTs^272,303  addressing this research question. We identified 40 studies^304-343  in these reviews that fulfilled our inclusion criteria and measured outcomes relevant to this context. Our systematic search for RCTs identified 6 additional studies^295,344-348  that fulfilled the inclusion criteria. Thirty studies reported the effect of LMWH vs UFH on risk of mortality.^{305,307,308,310-312,314,317-319,321,322,325-327,329-339,341,342,346,347}  Thirty-one studies reported the effect of LMWH vs UFH on development of symptomatic PEs,^{295,305-306,310-312,314-319,321,322,324,326-335,337,339,344-346}  5 studies reported the effect on symptomatic proximal DVTs,^{306,316,326,334,336}  and 7 reported the effect on symptomatic distal DVTs.^{306,316,326,329,334,336,341}  Thirty-four studies reported the effect of LMWH vs UFH on risk of major bleeding,^{295,305-307,311,313,315-319,321-339,341,344-346}  and 16 studies reported the effect on risk of reoperation.^{305,307,309,317,319,322,323,326,329,330,333-337,341} 

The EtD framework is available online at https://guidelines.gradepro.org/profile/EF7ADEA0-49F1-7E89-A0DB-DE7A9E854A2B.

Prophylaxis with LMWH vs UFH probably does not reduce mortality following major general surgery (RR, 1.03; 95% CI, 0.89-1.18; moderate certainty in the evidence of effects). We are very uncertain about the effect of LMWH on symptomatic PEs compared with that of UFH (RR, 0.83; 95% CI, 0.58-1.19; very low certainty in the evidence of effects), which corresponds to 1 fewer (3 fewer to 2 more) symptomatic PE per 1000 patients undergoing major general surgery based on a baseline risk of 0.8% from observational data.⁷³  It also appears to result in little or no difference in symptomatic proximal DVTs (RR, 1.01; 95% CI, 0.20-5.00; very low certainty in the evidence of effects) or a reduction in symptomatic distal DVTs (RR, 1.01; 95% CI, 0.30-3.44; very low certainty in the evidence of effects). We are very uncertain about the last 2 findings.

Prophylaxis with LMWH vs UFH probably does not affect major bleeding (RR, 0.97; 95% CI, 0.78-1.20; moderate certainty in the evidence of effects). This corresponds to 0 fewer (3 fewer to 3 more) major bleeding events per 1000 patients undergoing major general surgery. LMWH probably results in little or no difference in reoperations (RR, 0.79; 95% CI, 0.57-1.08; moderate certainty in the evidence of effects), which corresponds to 3 fewer (6 fewer to 1 more) reoperations per 1000 patients undergoing major general surgery.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations and very serious inconsistency.

There was possibly important uncertainty or variability in how patients valued the outcomes. Potential costs and savings were deemed of negligible relevance, assuming only in-hospital short-term prophylaxis. There was probably no impact on equity, because both LMWH and UFH were thought to be acceptable and feasible to implement. If extended prophylaxis beyond hospital discharge is planned, LMWH may be preferable given its once-daily dosing. The panel thought that both treatment options are already widely used and there should be few issues with implementation.

The guideline panel judged that the net benefit did not favor LMWH or UFH prophylaxis for patients undergoing major general surgery. Based on a very low overall certainty in the evidence, the panel determined that the balance of effects did not favor LMWH or UFH. In light of the very low certainty in the evidence, further high-quality comparative studies, using appropriate clinical outcomes, would be of value to add more certainty to this recommendation. However, such comparative studies are not regarded as high priority at this time.

Question: Should pharmacological prophylaxis vs no pharmacological prophylaxis be used for patients undergoing laparoscopic cholecystectomy?

We identified 1 systematic review of RCTs addressing this research question.³⁰  We identified 2 studies^118,349  in that review that fulfilled our inclusion criteria and measured outcomes relevant to this context. Because of the relative paucity of studies on patients undergoing laparoscopic cholecystectomy, data across major general, major gynecological, and major urological procedures were pooled, and laparoscopic cholecystectomy-specific baseline risk estimates were applied.³⁵⁰  Five studies^{66,294,301,351,352}  were conducted on patients undergoing major gynecological surgery. Thirty-one^{49,122,125,131,275-279,281-293,295-300,302}  studies were conducted on patients undergoing major general surgery. Two studies^118,349  were conducted on patients undergoing laparoscopic cholecystectomy. Six studies^{38,280,353-356}  were conducted on patients undergoing urological surgery. Eighteen studies^{122,125,274-278,284,285,291,293,296,297,299,300,302,353,354}  reported the effect of pharmacological prophylaxis vs no pharmacological prophylaxis on risk of mortality, 16 studies^{38,118,125,273,276-279,285,286,289,293,295,354-356}  reported the effect on risk of symptomatic PEs, 6 studies^{49,125,278,285,287,290}  reported the effect on risk of screening-detected proximal DVTs, and 7 studies^{66,125,278,284,285,287,290}  reported the effect on risk of screening-detected distal DVTs. Fifteen studies^{118,125,273-275,277,282,285,287,293,295,297,298,349,354}  reported the effect of pharmacological prophylaxis vs no pharmacological prophylaxis on risk of major bleeding, and 6 studies^{38,118,275,282,295,354}  reported the effect on risk of reoperation.

The EtD framework is available online at https://guidelines.gradepro.org/profile/E753AE97-D04A-D35F-ABE1-F9CAB9461DD1.

Pharmacological prophylaxis probably results in a small unimportant reduction in overall mortality (RR, 0.75; 95% CI, 0.61-0.93; low certainty in the evidence of effects). This corresponds to 2 fewer (0-2 fewer) deaths per 1000 patients based on a baseline risk of 0.6% from observational data.³⁵⁷  It may not reduce symptomatic PEs following laparoscopic cholecystectomy (RR, 0.48; 95% CI, 0.26-0.88; very low in the evidence of effects), but we are very uncertain of this finding. Given the very low baseline risk of VTE events in this specific patient population,³⁵⁰  this would be expected to result in 0 fewer (0 fewer to 0 more) symptomatic PEs per 1000 patients. Similarly, pharmacological prophylaxis may not reduce symptomatic proximal DVTs (RR, 0.38; 95% CI, 0.14-1.00; very low certainty in the evidence of effects) or symptomatic distal DVTs (RR, 0.52; 95% CI, 0.31-0.87; very low certainty in the evidence of effects). Again, given the very low baseline risks, this would correspond to 0 fewer (0 fewer to 0 more) events per 1000 patients for both outcomes. We are very uncertain about the effects on symptomatic proximal DVTs and symptomatic distal DVTs.

Pharmacological prophylaxis may result in a small increase in major bleeding (RR, 1.24; 95% CI, 0.87-1.77; low certainty in the evidence of effects). This would be expected to result in 6 more (3 fewer to 20 more) major bleeds per 1000 patients. We are very uncertain whether pharmacological prophylaxis results in little or no difference in reoperation (RR, 0.93; 95% CI, 0.35-2.50; very low certainty in the evidence of effects); this corresponds to 1 fewer (8 fewer to 18 more) reoperation per 1000 patients.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, indirectness, and imprecision.

The panel based this recommendation on the trivial incremental benefits and the small increased risk of major bleeding associated with pharmacological prophylaxis. The panel judged that the potential benefits of pharmacological prophylaxis were outweighed by the small increased risk of major bleeding in average-risk patients undergoing laparoscopic cholecystectomy. This relates to the very low baseline risk of VTE for patients undergoing laparoscopic cholecystectomy. The panel discounted the mortality difference observed in this analysis as unlikely to relate to pharmacological prophylaxis, given the very low baseline risk of VTE. Patients with other risk factors for VTE (eg, history of VTE, thrombophilia, or malignancy) may benefit from pharmacological prophylaxis.

The guideline panel determined that potential undesirable effects of pharmacological prophylaxis, in particular major bleeding, outweighed its potential benefit for patients undergoing laparoscopic cholecystectomy. The panel acknowledges that the overall certainty in the evidence was low, in particular as a result of indirectness, with most of the trial data included in the analysis involving patients undergoing major surgical procedures. Further research into pharmacological prophylaxis following laparoscopic cholecystectomy was not regarded as high priority given the low baseline incidence of VTE complications in this patient population.

Question: Should pharmacological prophylaxis vs no pharmacological prophylaxis be used for patients undergoing major neurosurgical procedures?

We identified 1 systematic review³⁵⁸  that addressed this question. We identified 6 studies in this review that fulfilled our inclusion criteria and measured outcomes relevant to this context.^{68,123,128,359-361}  Our update of the systematic review identified 1 additional study that fulfilled the inclusion criteria.¹³⁵  We additionally searched for and identified 3 nonrandomized studies that informed this question.^362-364  All studies included patients undergoing neurosurgical procedures. Five studies assessed the effect of LMWH,^{68,128,130,135,362}  4 studies assessed the effect of UFH,^359,361-363  1 study assessed the effect of warfarin,¹²⁴  and 1 study assessed the effect of heparin-dihydroergotamine.³⁶⁰  Additionally, across the 10 studies, mechanical prophylaxis was used as a cointervention in 6 of the randomized studies^{68,123,128,130,135,359}  and in all 3 of the nonrandomized studies.^361-363  Supplement 6 presents the characteristics of all included studies.

Five RCTs^{68,104,128,130,359,362}  and 2 nonrandomized studies^361,363  reported the effect of any pharmacological thromboprophylaxis vs no pharmacological intervention on mortality, 3 RCTs^123,128,360  and 2 nonrandomized studies reported on development of PEs,^361,363  2 RCTs reported on screening-detected proximal DVTs,^128,130  and 1 study reported on development of screening-detected distal DVTs.¹²⁸  Seven RCTs^{68,123,128,130,135,359,360}  and 3 nonrandomized studies^361-363  reported risk of major bleeding, and 2 RCTs reported on risk of reoperation.^135,359 

The EtD framework is available online at https://guidelines.gradepro.org/profile/C5A1B92D-0E70-50BA-847C-0497617938F5.

Pharmacological prophylaxis does not appear to reduce mortality (RR, 1.27; 95% CI, 0.57-2.69; low certainty in the evidence of effects) when considering the body of evidence from RCTs, in which we have more confidence than the nonrandomized data. This corresponds to 9 more deaths (15 fewer to 65 more) per 1000 patients. Based on RCT evidence, pharmacological prophylaxis may result in little or no difference in symptomatic PEs (RR, 0.84; 95% CI, 0.03-27.42; very low certainty in the evidence of effects), but we are very uncertain of this finding. This corresponds to 0 fewer events (2 fewer to 53 more) per 1000 patients based on a baseline risk of 0.2% from observational data.³⁶⁴  The absolute effect size is similarly small when considering the body of evidence from nonrandomized trials. Pharmacological prophylaxis may result in a small, possibly unimportant, effect on symptomatic proximal DVTs (RR, 0.50; 95% CI, 0.30 to 0.84; low certainty in the evidence of effects) and symptomatic distal DVTs (RR, 0.60; 95% CI, 0.33-1.08; very low certainty in the evidence of effects). This would correspond to 6 fewer (2-8 fewer) symptomatic proximal DVTs and 0 fewer (0-1 fewer) symptomatic distal DVTs per 1000 patients undergoing major neurosurgical procedures, based on baseline risks from observational data of 1.2% and 0.1%, respectively.³⁶⁴ 

From RCTs, pharmacological prophylaxis may increase major bleeding (RR, 1.57; 95% CI, 0.70-3.50; low certainty in the evidence of effects), corresponding to 10 more (5 fewer to 43 more) events per 1000 patients undergoing major neurosurgical procedures. We are very uncertain of the effect of pharmacological prophylaxis on reoperation (RR, 0.43; 95% CI, 0.06-2.84; very low certainty in the evidence of effects), which would correspond to 18 fewer (29 fewer to 57 more) reoperations per 1000 patients.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations and very serious imprecision.

The panel noted that observed RR ratios from observational studies tended to view pharmacological prophylaxis more favorably than RCTs. The panel based its recommendation on RR ratios from meta-analysis of RCTs rather than observational studies, given the greater risk of bias with the latter studies. The panel noted that the small benefit of pharmacological prophylaxis for the prevention of DVT was based upon randomized controlled studies using screening venography to detect rates of asymptomatic DVTs. The panel rated the harms of major bleeding associated with pharmacological prophylaxis as moderate because of the potential for greater morbidity associated with surgical site bleeding following these procedures. The panel recognized that patients undergoing major neurosurgical procedures would routinely receive prophylaxis with mechanical methods. The panel acknowledged that pharmacological prophylaxis might still be warranted in a higher-risk subgroup of patients, such as those experiencing prolonged immobility following surgery. In addition, the panel acknowledged that pharmacological prophylaxis could be considered for patients undergoing neurosurgical procedures that carried a lower risk for major bleeding. It may also be considered for patients with persistent mobility restrictions after the bleeding risk subsides following surgery.

The panel judged, based on the available very low certainty evidence, that the expected net benefit favored no prophylaxis following major neurosurgical procures, because the potential small benefit of reducing VTE events was outweighed by the potential moderate increased risk of major bleeding. The panel also recognized that mechanical methods of thromboprophylaxis are commonly used in this patient population. There is a great need for the performance of large RCTs evaluating pharmacological prophylaxis following major neurosurgical procedures, and using clinically important end points, to add certainty to this recommendation. The panel acknowledges that the current recommendation may not reflect standard practice in some centers.

Question: If pharmacological prophylaxis is indicated, should LMWH vs UFH be used for patients undergoing major neurosurgical procedures?

We identified 1 systematic review³⁵⁸  that addressed this question. We identified 4 studies in this review that fulfilled our inclusion criteria and measured outcomes relevant to this context.^348,365-367  Our update of the systematic review identified 1 additional study that fulfilled the inclusion criteria.³⁶⁸  All studies included patients undergoing neurosurgical procedures. Five studies reported the effect of LMWH compared with that of UFH on development of mortality,^348,365-368  2 studies reported on the development of PEs,^365,366  1 study reported on the development of screening-detected proximal DVTs,³⁴⁸  and 1 study reported on screening-detected distal DVTs.³⁶⁶  Four studies reported risk of major bleeding^348,365-367  and 1 study reported on risk of reoperation.³⁶⁶ 

All participants wore compression stockings, with the exception of 1 study in which their use was not reported.³⁶⁸ 

The EtD framework is available online at https://guidelines.gradepro.org/profile/E9D1EF22-EEC9-560E-A0CC-9FD435188BBE.

Pharmacological prophylaxis with LMWH may result in little to no difference in mortality compared with UFH (RR, 0.34; 95% CI, 0.04-3.21; low certainty in the evidence of effects) following major neurosurgical procedures; this corresponds to 3 fewer (5 fewer to 11 more) deaths per 1000 patients. LMWH may result in little or no difference in symptomatic PEs (RR, 0.20; 95% CI, 0.01-4.03; low certainty in the evidence of effects) compared with UFH following major neurosurgical procedures. Based on a baseline risk of 0.2% from observational data,³⁶⁴  this corresponds to 2 fewer (2 fewer to 6 more) symptomatic PEs in a cohort of 1000 patients following major neurosurgical procedures. We are uncertain whether LMWH affects proximal DVTs (RR, 1.00; 95% CI, 0.14-6.91; very low certainty in the evidence of effects). Similarly, the impact of LMWH on symptomatic distal DVTs is very uncertain (RR, 0.33; 95% CI, 0.01-7.93; very low certainty in the evidence of effects). In absolute terms, this corresponds to 1 fewer (1 fewer to 7 more) symptomatic distal DVT per 1000 patients, based on a baseline risk of 0.2% from observational data.³⁶⁴ 

LMWH may result in little or no difference in major bleeding (RR, 0.76; 95% CI, 0.20-2.95; low certainty in the evidence of effects) compared with UFH, which corresponds to 5 fewer (18 fewer to 43 more) major bleeding events per 1000 patients. We were unable to assess the effect on reoperations.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations and very serious imprecision.

In formulating this recommendation, the panel formally acknowledges its prior conditional recommendation against pharmacological prophylaxis (Recommendation 19) for patients undergoing neurosurgical procedures. Nevertheless, if pharmacological prophylaxis is considered, the panel judged the desirable effects of LMWH over UFH as small and the undesirable effects as trivial. Considering the very low certainty in the evidence, and possibly important uncertainty about or variability in how much people value the main outcomes, the balance of effects favored LMWH. Equity, acceptability, and feasibility were not considered major factors.

In the subset of high-risk patients following major neurosurgical procedures for whom pharmacological prophylaxis is being considered, the ASH guideline panel judged the net benefit to favor LMWH over UFH. The panel recognizes this as being based upon very low certainty in the evidence.

The research priorities following major neurosurgical procedures are to better establish the benefits and risks of any pharmacological prophylaxis compared with no pharmacological prophylaxis. For patients considered at very high risk of postoperative VTE and at low bleeding risk, high-quality comparative studies of LMWH vs UFH using clinically important outcome measures would be of value.

Question: Should pharmacological prophylaxis vs no pharmacological prophylaxis be used for patients undergoing TURP?

We did not identify any systematic reviews of RCTs addressing this research question. Because of the paucity of studies related to TURP, data across all major general, urological, and gynecological surgical procedures were pooled, and TURP-specific baseline risk estimates were applied, where available. The evidence base to inform the relative effectiveness of pharmacological prophylaxis vs no pharmacological prophylaxis was comparable to that used to inform this question for patients undergoing laparoscopic cholecystectomy (see Recommendation 18). To determine the desirable and undesirable effects of prophylaxis in absolute terms, the baseline risks specific for TURP outcomes were drawn from a systematic review by Tikkinen et al³⁶⁹  and from observational studies.^370-372 

The EtD framework is available online at https://guidelines.gradepro.org/profile/05201A35-BCDA-9EFA-98CB-892C0AB72944.

Pharmacological prophylaxis may not reduce overall mortality compared with no pharmacological prophylaxis (RR, 0.75; 95% CI, 0.61 to 0.93; low certainty in the evidence of effects). This corresponds to 1 fewer (0-1 fewer) death per 1000 patients undergoing TURP. Pharmacological prophylaxis may not reduce symptomatic PEs (RR, 0.48; 95% CI, 0.26 to 0.88; low certainty in the evidence of effects), which corresponds to 0 fewer events in lower-risk patients and 0 fewer (0-1 fewer) events in higher-risk patients undergoing TURP. We are very uncertain of its effect on symptomatic proximal DVTs (RR, 0.38; 95% CI, 0.14-1.00; very low certainty in the evidence of effects), which would correspond to 1 fewer (0-1 fewer) symptomatic event in 1000 lower-risk patients or 3 fewer (0-5 fewer) events per 1000 higher-risk patients. We are very uncertain about its effect on symptomatic distal DVTs (RR, 0.52; 95% CI, 0.31-0.87; very low certainty in the evidence of effects). Based upon the very low baseline risk for patients undergoing TURP, this would correspond to 0 fewer symptomatic events per 1000 higher-risk patients.

The risk of major bleeding is probably slightly increased (RR, 1.24; 95% CI, 0.87-1.77; moderate certainty in the evidence of effects) with the use of pharmacological prophylaxis. This corresponds to 6 more (3 fewer to 20 more) major bleeding events per 1000 patients. We are very uncertain about reoperations (RR, 0.93; 95% CI, 0.35-2.50; very low certainty in the evidence of effects) related to pharmacological prophylaxis.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, indirectness, and imprecision.

The panel judged the desirable effects of pharmacological prophylaxis for patients undergoing TURP as trivial and the undesirable effects as small in magnitude. It further judged that there was possibly important uncertainty or variability in how much people value the main outcomes. Based primarily on the very low baseline risk of VTE following TURP, the panel judged that the balance of effects ultimately favored not using pharmacological prophylaxis. Pharmacological prophylaxis would also incur moderate costs and not be cost-effective.

The guideline panel suggests against pharmacological prophylaxis for patients undergoing TURP. Based on overall very low certainty in the evidence, the panel judged that the desirable effects of pharmacological prophylaxis were outweighed by the undesirable effects, specifically the increased risk of bleeding in this setting. Given the very low baseline risks of VTE following this procedure and the increasing use of alternative modalities to treat lower urinary tract symptoms attributed to benign prostatic hyperplasia,³⁷³  further RCTs conducted on patients undergoing TURP do not appear to be a major priority.

Question: If pharmacological prophylaxis is indicated, should LMWH vs UFH be used for patients undergoing TURP?

We did not identify any systematic reviews of RCTs addressing this research question. Because of the paucity of studies related to TURP, data across all major general, urological, and gynecological surgical procedures were pooled. Thirty-one studies reported the effect of LMWH vs UFH on risk of mortality,^{305,307,308,310-312,314,317-319,321,322,325-327,329-339,341,342,346,347,374}  and 36 studies reported on development of symptomatic PEs.^{295,305-307,310-312,314-319,321,322,324,326-335,337,339,344-346,374-378}  Six studies reported the effect of LMWH vs UFH on development of symptomatic proximal DVTs,^{306,316,326,334,336,375}  8 studies reported the effect on development of symptomatic distal DVTs,^{306,316,326,329,334,336,341,375}  31 studies reported on risk of major bleeding,^{295,305-307,311,313,315-319,321-339,341,344-346,374-377,379}  and 19 studies reported on risk of reoperation.^{305,307,309,317,319,322,323,326,329,330,333-337,341,375-377}  To determine the desirable and undesirable effects of prophylaxis in absolute terms, the baseline risks specific for TURP outcomes obtained from a systematic review by Tikkinen et al³⁶⁹  were applied to the corresponding pooled RRs.

The EtD framework is available online at https://guidelines.gradepro.org/profile/59BDE78B-362E-9573-980D-41A280D79D9E.

The risks of mortality may be similar for patients treated with LMWH and UFH (RR, 1.03; 95% CI, 0.89-1.18; very low certainty in the evidence of effects), but we are very uncertain of this finding. Given an assumed baseline risk of 0.2%,^370-372  this corresponds to 0 fewer (0 fewer to 0 more) deaths per 1000 patients. Similarly, there may be little difference for other outcomes, such as symptomatic PEs (RR, 0.84; 95% CI, 0.59-1.20; very low certainty in the evidence of effects), corresponding to 0 fewer (0 fewer to 0 more) events per 1000 patients. We are equally uncertain about the risk of symptomatic proximal DVTs (RR, 1.01; 95% CI, 0.20-5.0; very low certainty in the evidence of effects) and symptomatic severe distal DVTs (RR, 1.01; 95% CI, 0.30-3.44; very low certainty in the evidence of effects). Corresponding absolute effect size estimates are very small.

The risks of bleeding may be similar with LMWH and UFH (RR, 0.97; 95% CI, 0.78-1.20; low certainty in the evidence of effects), corresponding to 0 fewer (4 fewer to 3 more) major bleeding events per 1000 patients undergoing TURP. The use of LMWH does not appear to decrease the risk of reoperation (RR, 0.79; 95% CI, 0.57-1.08; low certainty in the evidence of effects) for patients undergoing TURP. When applying a TURP-specific baseline risk of 0.2%, this corresponded to 0 fewer (0-1 fewer) reoperations per 1000 patients undergoing TURP.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, indirectness, and very serious imprecision.

The panel rated the magnitude of the desirable effects and undesirable effects of using LMWH or UFH as trivial. They further determined that there was possibly important uncertainty and/or variability in how much patients value the main outcomes. Overall, the balance of effects did not favor LMWH or UFH, nor did cost-effectiveness or issues surrounding equity, acceptability, and feasibility, at least for inpatient prophylaxis. The panel recognized that this particular comparison applied only to patients undergoing TURP considered at very high risk for VTE (eg, patients with a history of VTE) in whom pharmacological prophylaxis might be considered.

Based on very low certainty in the evidence for effects, the guideline panel did not find a net benefit for using LMWH or UFH for patients undergoing TURP in whom pharmacological prophylaxis is indicated and suggests that either can be used. For most patients undergoing TURP, the panel recommended against the use of pharmacological prophylaxis (see Recommendation 21). No high-priority research needs were identified.

Question: Should pharmacological prophylaxis vs no pharmacological prophylaxis be used for patients undergoing radical prostatectomy?

Because of the paucity of RCTs specific to this setting, the evidence base to inform the relative effectiveness of pharmacological prophylaxis vs no pharmacological prophylaxis was comparable to that used to inform this question for patients undergoing TURP (see Recommendation 21); we pooled data across all surgical procedures and applied surgery-specific baseline risk estimates for radical prostatectomy drawn from a systematic review by Tikkinen et al.³⁸⁰ 

The EtD framework is available online at https://guidelines.gradepro.org/profile/F99386B2-4C08-3F36-8029-61B51278B574.

Pharmacological prophylaxis vs no prophylaxis may not reduce mortality (RR, 0.75; 95% CI, 0.61 to 0.93; low certainty in the evidence of effects), corresponding to 0 fewer deaths per 1000 patients. Pharmacological prophylaxis probably does not reduce symptomatic PEs (RR, 0.48; 95% CI, 0.26-0.88; moderate certainty in the evidence of effects). Depending on patient risk, this corresponds to 0 fewer symptomatic PEs in lower-risk patients to 0 fewer (0-1 fewer) symptomatic PEs per 1000 higher-risk patients undergoing radical prostatectomy. We are very uncertain about the effect on symptomatic proximal DVTs (RR, 0.38; 95% CI, 0.14-1.00; very low certainty in the evidence of effects). This corresponds to 1 fewer (0-1 fewer) symptomatic proximal DVT per 1000 lower-risk patients to 4 fewer (0-5 fewer) per 1000 higher-risk patients. Distal DVTs may be reduced (RR, 0.52; 95% CI, 0.31-0.87; low certainty in the evidence of effects), but this also corresponds to a negligible effect of 0 fewer symptomatic distal DVT events, irrespective of baseline risk category.

Pharmacological prophylaxis compared with no prophylaxis may increase major bleeding (RR, 1.24; 95% CI, 0.87-1.77; low certainty in the evidence of effects). This corresponds to 6 more (3 fewer to 20 more) major bleeding events per 1000 patients undergoing radical prostatectomy. We are uncertain about the effect of pharmacological prophylaxis compared with no prophylaxis on reoperations (RR, 0.93; 95% CI, 0.35-2.50; very low certainty in the evidence of effects); this corresponds to 0 fewer (3 fewer to 6 more) reoperations per 1000 men based on a baseline risk of 0.4%.³⁸⁰ 

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, indirectness, and imprecision.

The panel recognized during its deliberations that the practice of radical prostatectomy varies greatly, ranging from robotically assisted laparoscopic radical prostatectomy with no or a limited pelvic lymph node dissection to open radical prostatectomy with extended pelvic lymph node dissection. Based on a systematic review by Tikkinen et al,³⁸⁰  this results in substantially different baseline risks for VTEs, which are important considerations for these recommendations. The majority of radical prostatectomies performed by urologists in the United States are performed robotically, typically with no or only a limited lymph node dissection. In this group, the panel judged the desirable effects of pharmacological prophylaxis as trivial and undesirable effects as small. Patients undergoing open radical prostatectomy with lymph node dissection were considered at higher risk for VTEs and bleeding. Cost-effectiveness probably favored no pharmacological prophylaxis, whereas issues of equity, acceptability, and feasibility were not deemed important in this setting.

The guideline panel judged that the net benefit favors no pharmacological prophylaxis for patients undergoing radical prostatectomy, based on very low certainty in the evidence of effects. The panel perceived it as important to emphasize that this recommendation was based on the panel’s assessment of average patients undergoing radical prostatectomy in the form of robotically assisted laparoscopic prostatectomy with no or limited lymph node dissection. Patients undergoing an extended node dissection and/or open radical prostatectomy may have a higher VTE risk and may potentially benefit from pharmacological prophylaxis. Further high-quality comparative studies, using appropriate clinical outcomes, would be of value to add more certainty to these recommendations. Further studies on patient values regarding prevention of VTEs and bleeding would allow for optimal shared decision making regarding thromboprophylaxis for radical prostatectomy.

Question: If pharmacological prophylaxis is indicated, should LMWH vs UFH be used for patients undergoing radical prostatectomy?

In the absence of RCTs specific to this setting, the evidence base to inform the relative effectiveness of LMWH prophylaxis vs UFH prophylaxis was comparable to that used to inform this question for patients undergoing TURP (see Recommendation 22); we pooled data across all surgical procedures and applied surgery-specific baseline risk estimates for radical prostatectomy drawn from a systematic review by Tikkinen et al.³⁸⁰ 

The EtD framework is available online at https://guidelines.gradepro.org/profile/AEF71CF4-AB9F-1DDF-A08B-1F5E2484EA5F.

The risks of mortality may be similar for patients treated with LMWH and UFH (RR, 1.03; 95% CI, 0.89-1.18; low certainty in the evidence of effects); this corresponds to 0 fewer deaths per 1000 men. Similarly, there may be little difference for the other outcomes of symptomatic PEs (RR, 0.84; 95% CI, 0.59-1.20; very low certainty in the evidence of effects), symptomatic proximal DVTs (RR, 1.01; 95% CI, 0.20-5.0; very low certainty in the evidence of effects), and symptomatic distal DVTs (RR, 1.01; 95% CI, 0.30-3.44; very low certainty in the evidence of effects), with all 95% CIs crossing the line of no effect and very small absolute effect sizes.

The risks of bleeding may be similar with LMWH and UFH (RR, 0.97; 95% CI, 0.78-1.20; low certainty in the evidence of effects), corresponding to 0 fewer (4 fewer to 3 more) major bleeding events per 1000 patients undergoing radical prostatectomy. The risks of reoperation may be similar with LMWH and UFH (RR, 0.79; 95% CI, 0.57-1.08; low certainty in the evidence of effects), corresponding to 1 fewer (0-2 fewer) event based on a baseline risk of 0.4%.³⁸⁰ 

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, indirectness, and very serious imprecision.

The panel rated the magnitude of the desirable and undesirable effects of using LMWH over UFH as trivial. They further determined that there was possibly important uncertainty and/or variability in how much people value the main outcomes. Overall, the balance of effects did not favor LMWH or UFH, nor did cost-effectiveness or issues surrounding equity, acceptability, and feasibility, at least for inpatient prophylaxis. The panel recognized that this particular comparison applied only to select patients undergoing radical prostatectomy considered at high risk for VTEs (eg, patients with prior VTEs). For most patients undergoing radical prostatectomy, the panel recommended against the use of pharmacological prophylaxis (see Recommendation 23).

The guideline panel judged that, for patients undergoing radical prostatectomy requiring pharmacological prophylaxis, based upon very low certainty in the evidence, LMWH or UFH can be used. There is a need for high-quality randomized trials specific to patients undergoing radical prostatectomy, particularly those treated with robotically assisted laparoscopic prostatectomy, the most widely used surgical approach for clinically localized prostate cancer.

Question: Should pharmacological prophylaxis vs no pharmacological prophylaxis be used for patients undergoing cardiac or major vascular surgery?

Our systematic search for RCTs identified 3 studies^70,381,382  that fulfilled our inclusion criteria and measured outcomes relevant to this context. Because of the paucity of RCTs, we also systematically searched for observational studies and identified 1 additional study that fulfilled our inclusion criteria.³⁸³ 

One trial reported the effect of pharmacological prophylaxis vs no intervention on risk of mortality.³⁸¹  Two trials reported the effect on the risk of symptomatic PEs and on the risk of any proximal and distal DVTs.^70,381,382  The effect on the risk of major bleeding was reported from an RCT³⁸¹  and a nonrandomized controlled study.³⁸³  Surgery-specific baseline risk estimates were obtained from a systematic review and meta-analysis of risk of VTE after cardiac surgery.³⁸⁴ 

The EtD framework is available online at https://guidelines.gradepro.org/profile/D5319730-F947-FFB7-B1F7-D4E9E4697079.

We were unable to assess the effect of pharmacological prophylaxis on mortality (RR, not estimable because no deaths were observed, low certainty in the evidence of effects). Pharmacological prophylaxis appears to result in little or no difference in symptomatic PEs (RR, 2.40; 95% CI, 0.10-55.7; low certainty in the evidence of effects); this corresponds to 5 more (3 fewer to 198 more) PEs per 1000 patients receiving pharmacological prophylaxis based on a baseline risk of 0.4%.³⁸⁴  We are very uncertain whether pharmacological prophylaxis results in little or no difference in proximal DVTs (RR, 2.85; 95% CI, 0.12-67.83; low certainty in the evidence of effects). This finding corresponds to 45 more (21 fewer to 1631 more) symptomatic proximal DVTs per 1000 patients, based on a baseline risk of 2.4%.³⁸⁴  We are also very uncertain about the effect of pharmacological prophylaxis on distal DVTs (RR, 0.32; 95% CI, 0.01-7.54; very low certainty in the evidence of effects). This corresponds to 1 fewer (2 fewer to 13 more) symptomatic distal DVT per 1000 high-risk patients, based on a baseline risk of 0.2%.³⁸⁴ 

Based on 1 large observational study³⁸³  and supported by a single relevant RCT,³⁸¹  the rates of major bleeding may be increased with pharmacological prophylaxis (RR, 1.26; 95% CI, 1.07-1.47; low certainty in the evidence of effects). This corresponds to 4 more (1-6 more) major bleeds per 1000 patients.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes informed by observational studies.

The panel recognized that very high doses of UFH are routinely administered to most patients undergoing cardiac and major vascular surgery. Thus, for this recommendation, the benefits and harms of postoperative pharmacological prophylaxis are being considered in an incremental context. The panel recognized there was a paucity of high-quality evidence addressing this particular question, and this recommendation was made in the face of very uncertain evidence. Based upon available evidence, the panel judged that the incremental desirable and undesirable effects of pharmacological prophylaxis were trivial and, therefore, balanced.

The panel recognized that cardiac surgery itself is associated with a risk for the development of heparin-induced thrombocytopenia (HIT). The panel reviewed the available literature and found that the risk of HIT among heparin preparations was higher with the use of UFH than with LMWH. Available evidence from RCTs did not allow the panel to quantitate whether there was an incremental risk for HIT associated with the use of pharmacological heparin prophylaxis beyond that of UFH administered during the procedure itself. However, the panel acknowledged that concerns about HIT would lead some panelists to be less likely to routinely use postoperative pharmacological prophylaxis with a heparin preparation, particularly UFH. The panel did not believe that there were important implementation considerations with the use of postoperative pharmacological prophylaxis in this patient population. Should LMWH or UFH be given, the panel recommended periodic monitoring of platelet counts because of the concern for postoperative HIT.

The panel found that the overall net benefit did not favor pharmacological prophylaxis or no pharmacological prophylaxis for patients undergoing cardiac and major vascular surgery; this was based on very low certainty in the evidence. For subgroup considerations, the panel judged that, for patients at higher baseline risk for VTE (eg, those with a history of VTE), pharmacological prophylaxis might be considered over no prophylaxis.

The panel supported that further research, in the form of well-designed RCTs using clinically important end points, is needed to determine the role of pharmacological prophylaxis in the prevention of VTEs following cardiac and major vascular surgery. Further research on the incremental impact of postoperative UFH and LMWH exposure on the development of HIT in this patient population is also warranted.

Question: Should LMWH prophylaxis vs UFH prophylaxis be used for patients undergoing cardiac or major vascular surgery?

We identified 1 systematic review³⁸⁵  of RCTs and observational studies that addressed this research question. We identified 3 studies^386-388  in these reviews that fulfilled our inclusion criteria and measured outcomes relevant to this context. Our update of the systematic review did not identify any additional studies that fulfilled the inclusion criteria. One study reported the effect of LMWH vs UFH on the risk of mortality, PE, and major bleeding.³⁸⁷  Three studies reported the effect of LMWH vs UFH on the risk of any DVT, including 1 study reporting data for proximal and distal DVTs separately.^386-398 

Surgery-specific baseline risk estimates were obtained from a systematic review and meta-analysis of risk of VTEs after cardiac surgery.³⁸⁴ 

The EtD framework is available online at https://guidelines.gradepro.org/profile/D32EF371-AE1E-1ACF-82CD-E11528E7B8E0.

We are very uncertain whether LMWH vs UFH affects mortality following cardiac or major vascular surgery (RR, 4.55; 95% CI, 0.22-93.81; low certainty in the evidence of effects). This corresponds to 0 fewer deaths per 1000 patients. We were unable to estimate an effect on symptomatic PEs (RR, not estimable). We are also very uncertain about the effect of LMWH on symptomatic proximal DVTs (RR, 1.33; 95% CI, 0.30-6.01; very low certainty in the evidence of effects) and symptomatic distal DVTs (RR, 1.20; 95% CI, 0.45-3.22; very low certainty in the evidence of effects). This would correspond to 8 more (17 fewer to 122 more) symptomatic proximal DVTs per 1000 patients and 0 more (1 fewer to 4 more) symptomatic distal DVTs per 1000 patients based on baseline risks of 2.4% and 0.2%, respectively.

LMWH vs UFH appears to result in little or no difference in major bleeding (RR, 0.91; 95% CI, 0.19-4.42; low certainty in the evidence of effects). We found no data on reoperation.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for indirectness and very serious imprecision.

The panel recognizes that they have judged that evidence was insufficient and of very low quality to recommend for or against pharmacological prophylaxis following cardiac surgery. Nevertheless, particularly for patients considered at high risk for VTE (eg, those with history of VTEs), postoperative pharmacological prophylaxis would be considered for use in the cardiac and major vascular surgery settings by some panelists. This recommendation is relevant for patients considered at high risk for VTEs.

As discussed in the previous recommendation, HIT is a recognized complication in the cardiac and vascular surgery settings. Furthermore, it is recognized that the risk of HIT in other settings has been shown to be higher with the use of UFH vs LMWH. Available evidence from RCTs did not allow the panel to quantitate whether there was an incremental risk for HIT associated with the use of pharmacological UFH or LMWH prophylaxis beyond that of heparin administered during the procedure itself or whether there was a relatively greater incremental risk for HIT in the cardiac surgery setting with postoperative UFH prophylaxis than with LMWH. Nevertheless, given the above factors, if pharmacological prophylaxis is chosen for use, panelists concurred that an anticoagulant with a lower risk for HIT (eg, LMWH over UFH) should be considered.

The panel did not believe that there were important implementation considerations with use of postoperative LMWH or UFH pharmacological prophylaxis in this patient population. Should either agent be given, the panel recommended the periodic monitoring of platelet counts.

The panel judged that, based upon available evidence, LMWH or UFH could be selected for VTE prophylaxis following cardiac surgery based on very low quality evidence. The panel judged that this question was only of relevance for patients considered at very high risk for postoperative VTEs following cardiac or major vascular surgery in whom pharmacological prophylaxis would be considered over no prophylaxis.

The panel supported that the more important research question for this patient population is the role of pharmacological prophylaxis vs no pharmacological prophylaxis for the prevention of VTEs following cardiac and major vascular surgery. Further research on the incremental impact of postoperative UFH and LMWH exposure on the development of HIT in this patient population would also be of value.

Question: Should pharmacological prophylaxis vs no pharmacological prophylaxis be used for patients experiencing major trauma?

We identified 1 systematic review that addressed this question.³⁸⁹  We identified 2 studies^47,390  in this review that fulfilled our inclusion criteria and measured outcomes relevant to this context. Our update of the systematic review did not identify any additional studies that fulfilled the inclusion criteria. Two small studies reported the effect of pharmacological prophylaxis compared with no intervention on risk of mortality, on development of symptomatic PEs, and on any DVT.^47,390  No study reported the effect of pharmacological prophylaxis compared with no intervention on risk of major bleeding or on risk of reoperation. The small amount of direct evidence, with a lack of information on undesirable outcomes, together with the very low certainty on the treatment effect, led the panel to consider the indirect data from hip fracture repair studies for treatment RR estimates^255-266  and applying baseline VTE and bleeding risks from studies on trauma patients.^391,392 

The EtD framework is available online at https://guidelines.gradepro.org/profile/434A9C2D-3417-F3ED-B7C1-4A0BA3EC6699.

We are uncertain about the effect of pharmacological prophylaxis on mortality following major trauma (RR, 0.95; 95% CI, 0.84-1.07; very low certainty in the evidence of effects). Pharmacological prophylaxis may reduce symptomatic PEs, but we are very uncertain of this finding (RR, 0.49; 95% CI, 0.33-0.72; very low certainty in the evidence of effects). Depending upon baseline risk, this benefit corresponds to 3 fewer (2-5 fewer) patients with symptomatic PEs per 1000 moderate-risk patients and 2 fewer (1-2 fewer) patients per 1000 low-risk patients.

Pharmacological prophylaxis may also reduce the risk of proximal DVTs (RR, 0.51; 95% CI, 0.38-0.69; very low certainty in the evidence of effects), which corresponds to 7 fewer (4-9 fewer) in 1000 higher-risk patients and 3 fewer (2-4 fewer) in 1000 lower-risk patients. We are uncertain of the effect of pharmacological prophylaxis on distal DVTs (RR, 0.85; 95% CI, 0.5-1.29, very low certainty in the evidence of effects). This would correspond to 1 fewer (3 fewer to 2 more) symptomatic distal DVT in 1000 higher-risk patients and 0 fewer in 1000 lower-risk patients.

Pharmacological prophylaxis may result in more major bleeding than no prophylaxis (RR, 1.24; 95% CI, 1.12-1.37; very low certainty in the evidence of effects), but this finding is uncertain. Depending upon baseline risk, the risk of major bleeding corresponds to 3 more (2-5 more) major bleeds per 1000 lower-bleeding-risk patients and to 14 more (7-21 more) major bleeds per 1000 patients with higher bleeding risk.

We are very uncertain about the effect of pharmacological prophylaxis on the need for reoperation (RR, 1.05; 95% CI, 0.82-1.35; very low certainty in the evidence of effects).

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, indirectness, and imprecision.

The overall certainty in the evidence was rated as very low for this question, given the absence of RCTs comparing pharmacological prophylaxis vs no prophylaxis for patients experiencing major trauma. The benefits and harm/burden data were extrapolated from the closest surgical indication for which we had adequate comparative evidence (ie, hip fracture repair). It is recognized that these hip fracture studies are dated and that rates of patient-important DVT events were derived from modeling of asymptomatic events detected by routine screening studies. The panel presumed that, in the absence of specific contraindications (eg, lower limb injuries), patients experiencing major trauma would receive mechanical prophylaxis. The panel emphasized the need to periodically reevaluate bleeding risk as patients recover from major trauma. Once bleeding is stabilized and the patient is no longer considered at high risk for major bleeding, the use of pharmacological prophylaxis should be reconsidered.

Based upon the totality of the evidence, the panel judged that the moderate overall benefits of pharmacological prophylaxis outweighed the increased risk of major bleeding for patients at low or moderate risk for bleeding. In contrast, for patients at high risk for major bleeding, the large undesirable consequences of major bleeding led to a balance that favors no pharmacological prophylaxis. The very low certainty in the evidence justifies conditional recommendations for both scenarios.

Well-designed trials using clinically important VTE end points are required for patients at low to moderate risk for bleeding following trauma to determine the incremental benefits of pharmacological prophylaxis beyond mechanical methods alone. Well-designed studies are also needed to determine the benefits and risks of introducing delayed pharmacological prophylaxis for patients experiencing major bleeding, including intracranial hemorrhage as a consequence of major trauma, as the bleeding risk subsides.

Question: Should LMWH vs UFH be used for patients experiencing major trauma?

We identified 1 systematic review that addressed this question.³⁸⁹  Our update of the systematic review identified 1 additional study that fulfilled the inclusion criteria,³⁹⁴  and our systematic search of RCTs identified 2 studies that fulfilled the inclusion criteria^394,395  and were not included in the review. Three studies reported the effect of LMWH prophylaxis vs UFH prophylaxis on risk of mortality, on development of any PEs, and on major bleeding,^393-395  whereas 2 studies informed on the risk of development of proximal and distal DVTs.^393,394 

The EtD framework is available online at https://guidelines.gradepro.org/profile/96D5A309-8606-4469-B732-E1844465CC75.

LMWH vs UFH appears to result in little or no difference in mortality for patients experiencing major trauma (RR, 1.32; 95% CI, 0.14-12.39; low certainty in the evidence of effects). This corresponds to 2 more (4 fewer to 54 more) deaths per 1000 trauma patients receiving LMWH vs UFH. Similarly, LMWH may result in little or no difference in symptomatic PEs (RR, 1.04; 95% CI, 0.11-9.92; low certainty in the evidence of effects). This corresponds to 0 fewer (6 fewer to 61 more) symptomatic PEs per 1000 patients based on a baseline risk of 0.7% from observational data.³⁹¹  LMWH vs UFH likely results in no important effect on proximal DVTs (RR, 0.57; 95% CI, 0.25-1.31; moderate certainty in the evidence of effects). This corresponds to 3 fewer (5 fewer to 2 more) symptomatic proximal DVTs per 1000 patients based on a baseline risk of 0.7% from observational data.³⁹¹ 

LMWH probably also results in little or no difference in symptomatic distal DVTs (RR, 0.74; 95% CI, 0.46-1.20; moderate certainty in the evidence of effects). This corresponds to 0 fewer symptomatic distal DVTs per 1000 patients based on a baseline risk of 0.1% from observational data.³⁹¹ 

LMWH may result in a small increase in major bleeding (RR, 2.4; 95% CI, 0.53-10.78; low certainty in the evidence of effects). This corresponds to 20 more (7 fewer to 138 more) major bleeding events per 1000 patients receiving LMWH vs UFH. There were no comparative data about rates of reoperation of LMWH vs UFH following major trauma.

We rated the overall certainty in the evidence of effects as low based on the lowest certainty in the evidence for the critical outcomes, downgrading for very serious imprecision.

The panel judged that the minor differences in the effects of the interventions on benefits and undesirable outcomes led to an overall balance that did not favor either intervention. The benefits observed with LMWH were limited to a minor reduction in the rates of symptomatic proximal DVTs, which was negated by the small observed increased risk of major bleeding. It was recognized that patients at high risk for major bleeding were excluded from the studies that formed the basis of this recommendation.

Because LMWH and UFH are in widespread use for this indication, the panel did not judge there to be major implementation considerations with either intervention. The panel presumed that, in the absence of specific contraindications (eg, lower limb injuries), patients experiencing major trauma would receive mechanical prophylaxis.

The panel suggested that, for patients experiencing major trauma who are judged to be at low to moderate risk for bleeding, LMWH or UFH may be used for pharmacological VTE prophylaxis based on low certainty in the evidence.

The panel judged that the research priorities in major trauma related to establishing the effectiveness and the timing of intervention with pharmacological prophylaxis for patients receiving mechanical prophylaxis following major trauma, rather than comparative studies of LMWH vs UFH.

Question: Should pharmacological prophylaxis vs no pharmacological prophylaxis be used for patients undergoing major gynecological surgery?

We identified 1 systematic review of RCTs addressing this research question.²⁹  We identified only 5 studies^{118,273-275,349}  in this review that fulfilled our inclusion criteria and measured outcomes relevant to this context. Because of the paucity of studies on patients undergoing major gynecological procedures, data across major general, laparoscopic cholecystectomy, and major gynecological and urological procedures were pooled. The evidence base to inform the relative effectiveness of pharmacological prophylaxis vs no pharmacological prophylaxis was comparable to that used to inform this question for patients undergoing laparoscopic cholecystectomy (see Recommendation 18). Baseline risk estimates specific to gynecological procedures^396,397  were applied to determine the desirable and undesirable effects of prophylaxis in absolute terms.

The EtD framework is available online at https://guidelines.gradepro.org/profile/B2FDFE66-5A79-4E46-875E-9BB7F3FAFF9F.

Pharmacological prophylaxis probably reduces mortality slightly following major gynecological surgery (RR, 0.75; 95% CI, 0.61-0.93; low certainty in the evidence of effects). This corresponds to a benefit of 4 fewer (1-7 fewer) deaths per 1000 patients. Pharmacological prophylaxis probably reduces the risk of symptomatic PEs (RR, 0.48; 95% CI, 0.26-0.88; low certainty in the evidence of effects), corresponding to a benefit of 2 fewer (0-3 fewer) symptomatic PEs per 1000 higher-risk patients and 0 fewer (0-1 fewer) per 1000 lower-risk patients, based on baseline risks from observation data of 0.1% and 0.4%, respectively.^396,397 

We are uncertain whether pharmacological prophylaxis reduces the rates of proximal DVTs (RR, 0.38; 95% CI, 0.14-1.00; very low certainty in the evidence of effects), corresponding to reduction of 4 fewer (0-6 fewer) symptomatic proximal DVTs per 1000 lower-risk patients and 17 fewer (0-23 fewer) per 1000 higher-risk patients based on baseline risks from observation data of 0.7% and 2.7%, respectively.^396,397  We are uncertain about the effect of pharmacological prophylaxis on distal DVTs (RR, 0.52; 95% CI, 0.31-0.87; low certainty in the evidence of effects), corresponding to no reduction per 1000 patients treated based on a lower baseline risk of 0.1%.

Pharmacological prophylaxis may slightly increase the risk of major bleeding (RR, 1.24; 95% CI, 0.87-1.77; low certainty in the evidence of effects). This corresponds to 6 more (3 fewer to 20 more) major bleeding events per 1000 patients receiving pharmacological prophylaxis. Pharmacological prophylaxis does not appear to increase the risk of reoperation (RR, 0.93; 95% CI, 0.35-2.50; very low certainty in the evidence of effects) following major gynecological procedures.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations, indirectness, and imprecision.

The panel based its recommendation on the judgment that the desirable benefits of pharmacological prophylaxis outweighed the likely small increased risk of major bleeding following major gynecological procedures. The panel acknowledges that the overall certainty in the evidence was very low given the issue of indirectness, with most of the available trial data not being specific to gynecological procedures. The panel considered that patients at increased risk for VTE would receive mechanical prophylaxis in addition to pharmacological prophylaxis. There were no major implementation considerations.

The panel judged that pharmacological prophylaxis should be administered to patients undergoing major gynecological surgery, and this recommendation was conditional given the very low certainty in the evidence. There is a need for large high-quality clinical trials using clinically relevant end points to determine the benefit of pharmacological prophylaxis following gynecological procedures. These studies should include detailed clinical characteristics of the patient populations.

Question: Should LMWH vs UFH prophylaxis be used for patients undergoing major gynecological surgery?

We identified 1 systematic review of RCTs addressing this research question.³⁰  We identified only 4 studies^{374,376,377,379}  overall that were conducted with patients undergoing major gynecological surgery. Because of the paucity of studies on patients undergoing major gynecological procedures, data across major general, laparoscopic cholecystectomy, and major gynecological and urological procedures were pooled. The evidence base to inform the relative effectiveness of LMWH prophylaxis vs UFH prophylaxis was comparable to that used to inform this question for patients undergoing TURP (see Recommendation 22).

Baseline risk estimates specific to gynecological procedures^396,397  were applied to determine the desirable and undesirable effects of prophylaxis in absolute terms.

The EtD framework is available online at https://guidelines.gradepro.org/profile/F213C6D1-F2D9-221A-B8EE-92B6F94F5BB3.

LMWH prophylaxis appears to result in little or no difference in mortality compared with UFH prophylaxis (RR, 1.03; 95% CI, 0.89-1.18; low certainty in the evidence of effects), corresponding to 1 more (2 fewer to 3 more) deaths per 1000 patients. Likewise, use of LMWH prophylaxis vs UFH prophylaxis appears to result in little or no difference in symptomatic PEs (RR, 0.91; 95% CI, 0.63-1.3; low certainty in the evidence of effects). For a higher baseline risk of 0.4% from observational data,³⁹⁶  this corresponds to 1 fewer (2 fewer to 1 more) symptomatic PE per 1000 participants. For symptomatic proximal DVTs (RR, 1.01; 95% CI, 0.20-5.00; low certainty in the evidence of effects), the absolute risk reduction is 0 per 1000 patients, with the 95% CI varying by baseline risk from 5 fewer to 27 more³⁹⁷  to 22 fewer to 108 more (baseline risks of 0.7% and 2.7%, respectively).³⁹⁶  We are very uncertain about the effect on symptomatic distal DVTs (RR, 1.01; 95% CI, 0.30-3.44; very low certainty in the evidence of effects) following major gynecological surgical procedures.

LMWH appears to confer little or no difference in major bleeding compared with UFH prophylaxis following major gynecological procedures (RR, 0.97; 95% CI, 0.78-1.20; low certainty in the evidence of effects); this corresponds to 0 fewer (4 fewer to 3 more) major bleeds per 1000 patients. LMWH prophylaxis may result in a small, possibly unimportant, reduction in reoperations compared with UFH (RR, 0.79; 95% CI, 0.57-1.08; low certainty in the evidence of effects). This corresponds to 4 fewer (8 fewer to 1 more) reoperation procedures per 1000 patients receiving LMWH prophylaxis.

We rated the overall certainty in the evidence of effects as very low based on the lowest certainty in the evidence for the critical outcomes, downgrading for study limitations and indirectness.

The panel based its recommendation on the judgment that the desirable benefits and the risk of complications were balanced between use of LMWH and UFH pharmacological prophylaxis. The panel acknowledges that the overall certainty in the evidence was very low, given the issue of indirectness, with most of the available trial data not being specific to gynecological procedures. There was not deemed to be any major implementation consideration.

The panel suggests using LMWH or UFH for patients undergoing major gynecological surgery procedures based upon very low certainty in the evidence of effects. There is a need for large high-quality clinical trials using clinically relevant end points to determine the relative benefits of LMWH vs UFH pharmacological prophylaxis following gynecological procedures. These studies should include detailed clinical characteristics of the patient populations.

The panel recognized that many studies of pharmacological and mechanical prophylaxis for VTE prevention following major surgery date back decades, thereby raising questions about the applicability of this evidence. This includes largely outdated means (eg, venography) to assess for VTEs postoperatively. Surgical practice has changed considerably over the decades, aimed at improving the patient experience. In most circumstances, these innovations would be expected to reduce the overall risk of postoperative VTEs. Examples of such innovations include use of minimally invasive surgical procedures, early and increased postoperative patient mobilization, and use of regional anesthesia; however, it is uncertain whether such changes in surgical practice impact the relative effectiveness of various thromboembolic interventions. Therefore, for the purpose of this guideline, this type of evidence informing the relative effectiveness of these interventions was included when obtained in the setting of RCTs. For determining baseline risk of VTEs and major bleeding, we used data, where available, from contemporary large cohort studies that were deemed representative of contemporary patients.

Although the panel rated symptomatic VTE end points as those upon which recommendations should be based, the panel recognized that most studies of VTE prophylaxis following surgery used asymptomatic DVTs detected by the routine performance of sensitive screening tests (eg, venography) as the primary study outcome. Reporting of symptomatic DVTs in some studies could have been influenced by diagnostic suspicion bias. The panel also acknowledges that modeling was required to determine rates of symptomatic DVTs when only asymptomatic DVT events were reported, based on the best available estimates drawn from the literature.

The panel recognized that most of the evidence on mechanical methods of VTE prophylaxis comes from the orthopedic literature. Studies of the benefits of mechanical prophylaxis for other surgical settings are needed. Finally, the panel acknowledges that, for some questions, limited direct data were available (eg, VTE prophylaxis following urological and gynecological procedures and for major trauma). In these settings, estimates of the benefits of prophylaxis were based upon related surgical settings, such as general surgical procedures and hip fracture surgery, respectively.

These ASH guidelines stand out by their scope, which includes general issues relevant to any surgical procedure and those related to surgical subspecialties. They make consistent use of high-quality systematic reviews and provide a formal EtD framework for every recommendation, thereby enhancing transparency about the judgments that were made.

A widely used high-quality guideline is the 2012 Guideline of the American College of Chest Physicians (ACCP), which places a strong emphasis on patients’ VTE risk scores.³⁹⁸  In the guideline recommendations for VTE prevention in nonorthopedic surgical patients, patient-oriented VTE risk calculators, such as the Caprini score¹⁰  and Rogers score,³⁹⁹  were adopted. When the risk for VTE is very low, it was recommended not to use pharmacological or mechanical prophylaxis. For patients at low risk for VTE, mechanical prophylaxis was suggested over no prophylaxis, preferably with intermittent pneumatic compression. For patients at moderate risk for VTE who are not at high risk for major bleeding complications, it was suggested to use LMWH, low-dose UFH, or mechanical prophylaxis with intermittent pneumatic compression over no prophylaxis. For patients at high risk for VTE who are not at high risk for major bleeding complications, it was recommended to use pharmacological prophylaxis with LMWH or low-dose UFH over no prophylaxis, and it was suggested to add mechanical prophylaxis with graduated compression stockings or intermittent pneumatic compression to pharmacological prophylaxis. Guidelines by the Scottish Intercollegiate Guidelines Network (SIGN) and the National Institute for Clinical Excellence (NICE) also emphasize the importance of risk stratification.^400,401 

Mechanical methods of perioperative VTE prophylaxis have been addressed by a number of guidelines.^398,400-404  The 2012 ACCP guidelines did not provide a detailed comparison of the effectiveness of graduated compression stockings and pneumatic compression stockings, but they generally favor pneumatic compression stockings on the basis of indirect evidence, from the Clots in Legs or Stockings after Stroke trial for patients with stroke, that elastic stockings increased the risk of skin complications without reducing the risk of VTEs.⁴⁰⁵  Other guidelines presented by SIGN,⁴⁰⁰  NICE,⁴⁰¹  the American Academy of Orthopedic Surgeons (AAOS),⁴⁰²  the Neurocritical Society,⁴⁰⁴  and the International Union of Angiology⁴⁰³  discuss the use of pneumatic compression and graduated compression stockings in various surgical settings but generally consider these modalities together as “mechanical devices” and do not offer a direct comparison of their effectiveness or safety other than noting that graduated compression stockings cannot be used for patients with certain lower extremity pathologies.

Guidelines addressing the prophylactic placement of IVC filters include the 2012 ACCP guidelines,³⁹⁸  the 2011 AAOS guideline for orthopedic patients,⁴⁰²  the 2013 European Venous Forum,⁴⁰³  the 2013 guidelines by the Neurocritical Care Society,⁴⁰⁴  the 2013 British Committee for Standards in Hematology guidelines, and the “appropriateness criteria” by the American College of Radiology.⁴⁰⁶  The recommendation made by these ASH guidelines corresponds with many of these existing recommendations that are mostly critical of prophylactic IVC filter placement for patients requiring major surgery or who have experienced trauma.

Several recent guidelines comment on VTE prophylaxis after total hip or knee arthroplasty. The 2012 ACCP guideline for orthopedic surgery patients⁴⁰⁷  recommended LMWH, fondaparinux, apixaban, dabigatran, rivaroxaban, low-dose UFH, adjusted-dose vitamin K antagonists (VKAs), ASA, and/or intermittent pneumatic compression, with the proviso that they are portable, out of concerns regarding compliance. They further indicate a preference for LMWH over the other listed agents, with the exception of ASA. The 2011 AAOS guideline⁴⁰²  recommends some form of chemoprophylaxis (including ASA) along with intermittent pneumatic compression after total hip or knee arthroplasty. The 2012 Asian Venous Thromboembolism Guideline⁴⁰⁸  recommends LMWH, fondaparinux, rivaroxaban, apixaban, edoxaban, dabigatran, warfarin, or ASA with intermittent pneumatic compression, referencing and effectively accepting the 2012 ACCP and 2011 AAOS guidelines. The 2013 International Angiology guideline favors LMWH, fondaparinux, VKAs, rivaroxaban, apixaban, or dabigatran, along with use of intermittent pneumatic compression after total hip arthroplasty.⁴⁰³  The most current NICE guideline recommends LMWH or rivaroxaban after total hip arthroplasty and the same after total knee arthroplasty, with the additional option of ASA.⁴⁰¹ 

For VTE prophylaxis after surgery for hip fractures, the 2012 ACCP guideline recommends LMWH for VTE prophylaxis vs fondaparinux and low-dose UFH over adjusted-dose VKAs or ASA.⁴⁰⁷  Concurrent use of an intermittent pneumatic compression device was also recommended. The 2011 AAOS guideline does not specifically address hip fractures; however, in the face of hip arthroplasty as a treatment for hip fracture, their recommendations of some form of chemoprophylaxis (including ASA), along with intermittent pneumatic compression for total hip arthroplasty, would be applicable.⁴⁰²  The 2012 Asian Venous Thromboembolism Guideline recommended LMWH, fondaparinux, rivaroxaban, apixaban, edoxaban, dabigatran, warfarin, or ASA with intermittent pneumatic compression.⁴⁰⁸  The Agency for Healthcare Research and Quality 2017 guideline favors chemoprophylaxis but is neutral on specific agents because of a lack of evidence.²³⁶  The 2013 International Angiology Guideline favors LMWH, fondaparinux, VKAs, or low-dose UFH.⁴⁰³  The most current NICE guideline recommends LMWH or fondaparinux.⁴⁰¹ 

For general and abdominal surgery, which includes gastrointestinal, urological, gynecological, bariatric, vascular, plastic, or reconstructive surgery in its scope, the 2012 ACCP guidelines are once again the best known. In the very low risk setting, no specific pharmacological or mechanical prophylaxis is recommended other than early ambulation. At low risk, mechanical prophylaxis (preferably with intermittent pneumatic compression) is suggested over no prophylaxis. For moderate-risk patients, assuming there is no high risk for major bleeding, LMWH, low-dose UFH, and mechanical prophylaxis, preferably with intermittent pneumatic compression devices, are all options. If patients are at high risk for major bleeding complications or if consequences of bleeding are thought to be particularly severe, mechanical prophylaxis (preferably with intermittent pneumatic compression devices) is suggested over no prophylaxis.³⁹⁸ 

For patients undergoing neurosurgical procedures, there are a total of 6 guidelines that offer somewhat conflicting recommendations on pharmacological VTE prophylaxis. These guidelines include the 2012 ACCP guideline, the 2014 Korean Society of Thrombosis and Hemostasis Evidence-Based Clinical Practice Guidelines, the 2018 NICE guideline, the 2010 SIGN guideline, the 2017 European Society of Anesthesiology guideline, and the 2016 Neurocritical Care Society guidelines.^{398,400,401,404,409,410}  The recommendations provided by these current ASH guidelines are similar to the 2010 SIGN guidelines, the 2019 Congress of Neurological Surgeons Guidelines for Spine Trauma, and the 2012 ACCP guidelines. Overall, mechanical prophylaxis is recommended for most neurosurgical patients. The benefit of pharmacological prophylaxis should be considered for patients at high risk or very high risk for VTE, but the overall risk/benefit profile is questionable, given an increase in bleeding events, particularly because neurosurgical bleeding events can be more serious. Similar to the 2010 SIGN recommendations, our ASH guidelines suggest that, for patients who do receive pharmacological prophylaxis, LMWH be used over UFH, whereas the 2012 ACCP guidelines do not give preference to any specific drug.

Urology is covered within the scope of guidelines by the ACCP, the Australian National Health and Medical Research Council, and the National Institute for Health and Care Excellence, among others.⁴¹¹  Urology-specific guidelines are available from the American Urological Association, the German Association of Scientific Medical Societies, and, most recently, the European Association of Urology.^412-414  In the absence of direct evidence for urology, the European Association of Urology guideline makes the assumption of a 50% risk reduction for “any serious VTE event” as well as a 50% increase in the risk of major bleeding requiring reoperation. When assessing the net benefit in making the recommendation, the major bleeding was given twice the weight of the outcome of VTE prevention. The resulting recommendations were supported by a systematic review of the procedure-specific VTE risk and the bleeding risk.^369,380  For patients undergoing TURP, this resulted in a conditional recommendation against pharmacological prophylaxis across risk groups. For radical prostatectomy, the guideline provides a more nuanced set of recommendations that differ by surgical approach (open, laparoscopic, or robotically assisted laparoscopic) and extent of the node dissection (without, standard, or extended). For patients undergoing robotically assisted laparoscopic prostatectomy with a standard lymph node dissection, which was considered the index case for these ASH guidelines, there is a conditional recommendation against pharmacological prophylaxis.

For cardiac surgery patients with an uncomplicated postoperative course, the 2012 ACCP guideline suggested the use of mechanical prophylaxis, preferably with optimally applied intermittent pneumatic compression, over no prophylaxis or pharmacological prophylaxis.³⁹⁸  For cardiac surgery patients whose hospital course is prolonged by ≥1 nonhemorrhagic surgical complication, the guideline suggested adding pharmacological prophylaxis with UFH or LMWH to mechanical prophylaxis. They considered that the risk of VTE following cardiac surgery is uncertain but judged that most patients were at moderate risk for VTEs and at high risk for anticoagulant prophylaxis-related bleeding. Based on these considerations, it was concluded that, in cardiac patients at usual risk for VTE, the harms of anticoagulant prophylaxis outweighed the benefits, whereas anticoagulation may be of net benefit for patients with high-risk characteristics. The ACCP guidelines did not provide recommendations specific to major vascular surgery. The International Consensus Statement on Prevention and Treatment of Venous Thromboembolism published by the European Venous Forum, in cooperation with several other organizations, offers guidelines for general, vascular, bariatric, and plastic surgical patients.⁴⁰³  Major vascular surgery was considered with other “major surgery,” and patients were judged to generally be at moderate risk in the absence of specific high-risk characteristics, such as age older than 60 years or prior VTE. Pharmacological prophylaxis was recommended in the absence of unusual bleeding risks. Recommendations specific to cardiac surgery patients were not presented. An update of NICE guidelines published in 2018 offers guidelines regarding VTE prophylaxis for patients undergoing cardiac or major vascular surgery.⁴⁰¹  This guideline recommends considering mechanical VTE prophylaxis on admission for patients who are undergoing cardiac surgery and are at increased risk for VTE and continuing this until the patient no longer has significantly reduced mobility relative to their normal or anticipated mobility. They recommend considering pharmacological VTE prophylaxis for a minimum of 7 days for patients who are undergoing open vascular surgery or major endovascular procedures, including endovascular aneurysm repair, and whose risk of VTE outweighs their risk of bleeding. Further, they recommend considering mechanical prophylaxis on admission for people who are undergoing open vascular surgery or major endovascular procedures, including endovascular aneurysm repair, if pharmacological prophylaxis is contraindicated. SIGN published a relevant updated guideline in 2014.⁴⁰⁰  The guideline notes that cardiac surgery patients often receive anticoagulants and antiplatelet agents for reasons independent of VTE and that this may impact their VTE risk. Despite this, these guidelines recommend that patients undergoing coronary artery bypass grafting surgery should be offered mechanical thromboprophylaxis where feasible and that patients undergoing coronary artery bypass grafting surgery who are not at high risk for bleeding can also be offered pharmacological thromboprophylaxis.

The most relevant guideline on the perioperative management of trauma patients is that by the ACCP in 2012.³⁹⁸  For major trauma, including traumatic brain injury, acute spinal injury, and traumatic spine injury, VTE prophylaxis is suggested (over no prophylaxis) with LMWH, low-dose UFH, or mechanical prophylaxis, preferably with intermittent pneumatic compression. For patients at high risk for VTE, addition of mechanical prophylaxis to pharmacological prophylaxis is suggested when not contraindicated by lower extremity injury. It further suggests against placement of an IVC filter for primary VTE prevention, as well as against periodic surveillance with venous compression ultrasound.

A guidance document from the American College of Gynecology dates back to 2007⁴¹⁵ ; as a result, the 2012 ACCP guidelines provide the timeliest guidance for gynecological surgery. One set of recommendations is made across gastrointestinal, urological, gynecological, bariatric, vascular, plastic, and reconstructive surgery (see above).

After publication of these guidelines, ASH will maintain them through surveillance for new evidence, ongoing review by experts, and regular revisions.

Adaptation of these guidelines will be necessary in many circumstances. These adaptations should be based on the associated EtD framework.⁴¹⁶  The Agency for Healthcare Research and Quality in the United States provides a guide for implementing effective quality improvement in this patient population.⁴¹⁷ 

The authors thank Joyce Kirkman and Tracy Minichiello for participation on the panel during the initial stages of the guideline-development process, including to prioritize questions and health outcomes. The authors acknowledge Angela Barbara, Housne Begum, Samantha Craigie, Andrea Darzi, Farid Foroutan, Nora Ibargoyen Roteta, Rana Qadeer, Yasir Rehman, Mansoor Radwi, Matthew Ventresca, and Robby Nieuwlaat for supporting the systematic reviews and Wojtek Wiercioch, Robby Nieuwlaat, and Holger Schünemann for coordination of the systematic review team. Holger Schünemann prepared a template of these guidelines for all panels that was critically reviewed by Adam Cuker, Rob Kunkle, the ASH Guideline Oversight Subcommittee, the Methods Group supporting the guidelines, and Blood Advances editors.

Contribution: D.R.A. and P.D. wrote the first draft of the manuscript and revised the manuscript based on the authors’ suggestions; C.B., F.D., C.W.F., D.A.G., S.R.K., M.R., A.R., F.B.R., M.A.S., K.A.O.T., and A.J.Y. critically reviewed the manuscript and provided suggestions for improvement; members of the knowledge synthesis team (G.P.M., T.B., S.B., J.L.B., I.E.-I., H.J., I.N., W.W., J.J.Y.-N., and H.J.S.) contributed evidence summaries to the guidelines; D.R.A. and P.D. were the chair and vice chair of the panel and led the panel meeting; and all authors approved the content.

Conflict-of-interest disclosure: All authors were members of the guideline panel or of the systematic review team. As such, they completed a disclosure of interest form, which was reviewed by ASH and is available as Supplements 2 and 3.

Correspondence: David R. Anderson, Faculty of Medicine, Dalhousie University, Clinical Research Centre, Room C-205, 5849 University Ave, PO Box 15000, Halifax, NS B3H 4R2, Canada; e-mail: david.anderson@dal.ca.