COMMENT|ONLINE FIRST

Improved precision in defining the need for thromboprophylaxis during lower limb immobilisation

Beverley J Hunt, Xavier L Griffin

Lancet Published: February 15, 2024

DOI:https://doi.org/10.1016/S0140-6736(23)02613-2

Hospital-associated venous thromboembolism is defined as venous thromboembolism occurring during hospital admission and up to 90 days after discharge. It is the major cause of venous thromboembolism, accounting for over 50% of all cases and is a global health issue, with risk of death and long-term morbidity, such as post-thrombotic syndrome, pulmonary hypertension, and psychological sequelae.¹ Globally, there are an estimated 10 million cases of hospital-associated venous thromboembolism.¹

Temporary lower limb immobilisation (eg, plaster cast, walking boot, or rigid splint) and injury are combined risk factors for hospital-associated venous thromboembolism (hazard ratio 6·31 [95% CI 5·30–7·52]).² The number of people affected is large; for example, approximately 70 000 people are immobilised after injury annually in the UK.³ The risk of hospital-associated venous thromboembolism in this setting is potentially reduced with pharmacological thromboprophylaxis. Currently baseline venous thromboembolism risk varies widely across the population and there are multiple drugs available for prophylaxis, including direct oral anticoagulants and low molecular weight heparin. However, there is no high-quality evidence comparing different anticoagulants in patients at high risk of venous thromboembolism; little evidence of which patients are at low risk; and the benefits and risks of any prophylaxis for those at low risk. Thus, there has been extensive variation in care. Furthermore, international guidelines offer conflicting advice, with some advocating pharmacological thromboprophylaxis for all, leading to increased risk of adverse events, notably bleeding and health-care costs.

The TiLLI research group explored the cost-effectiveness of risk assessment and different drug thromboprophylaxis of lower limb immobilisation in a systematic review, network meta-analysis, and decision analysis model,^4,  5 showing that drug prophylaxis for all patients had a 76% probability of being cost-effective across the UK National Health Service, at a threshold of £20 000 per quality-adjusted life-year. However, a sensitivity analysis showed that risk-based prophylaxis strategies could be more cost-effective, but only if an adequately validated and sensitive (84–89%) risk assessment tool were available.^{6,  7,  8,  9}

There is now a sophisticated and safe approach for venous thromboembolism risk assessment in lower limb immobilisation with the TRiP(cast) tool. Studies have shown the TRiP(cast) tool to be superior to other tools^{10,  11,  12,  13} and now it has been fully validated in Delphine Douillet and colleagues' completed stepped wedge, randomised trial,¹⁴ which included 2120 participants (49·8% were women, median age was 35 [IQR 24–49] years) across 15 hospitals in France and Belgium. The aim of the CASTING trial was to evaluate the safety profile of withholding pharmacological prophylaxis for individuals prescribed lower limb immobilisation for acute injury who were deemed at low absolute risk of venous thromboembolism defined by a TRiP(cast) of less than 7. Hospitals were randomly assigned to switch strategy from a control of prescribing pharmacological prophylaxis as per the physicians' usual practice to prescribing prophylaxis only to individuals with a score of at least 7. The primary outcome was the symptomatic venous thromboembolism rate (ie, the cumulative incidence of symptomatic deep vein thrombosis or pulmonary embolism, fatal pulmonary embolism, and unexplained sudden death) within 3 months after inclusion. There was a large and significant reduction in prescribing between phases (control prescription rate 50·4%; absolute difference –26·0 percentage points [95% CI –30·2 to –21·7, p value <0·0001]). Despite this marked reduction in prescribing, there was no difference between groups in the cumulative rate of symptomatic venous thromboembolism (control risk 1·0%, absolute difference 0·1% [95% CI –0·8 to 1·1]).

At this threshold, TRiP(cast) delivers the optimal trade-off between sensitivity and specificity that has been previously identified to be effective (sensitivity 85·7%, specificity 32·2%, negative predictive value 99·3%) and outperforms less personalised strategies. As a direct consequence of this study, patients at low risk of venous thromboembolism can safely avoid ineffective and potentially harmful anticoagulation.

Where do we need to go now? Validation of this tool allows health-care professionals to stratify the baseline risk for patients prescribed lower limb immobilisation with precision and design personalised research questions to subpopulations with different baseline venous thromboembolism risks. The most pressing question now is to determine the most effective and safe personalised anticoagulant for individuals. In the UK, a new randomised trial (NIHR154716), funded by the National Institute for Health and Care Research, will test oral anticoagulation, parenteral anticoagulation, and no drug, with participants stratified by baseline risk using the tool validated by Douillet and colleagues. The trial treatments will be defined by the TRiP(cast) risk assessment, such that we can further substantiate the CASTING trial results for individuals at low risk and compare different drug therapies for individuals at high risk. Meanwhile, Douillet and colleagues will start the RIVACAST randomised controlled trial (NCT06195540) comparing direct oral anticoagulants with low molecular weight heparin in patients at high risk. These trials are based on the validation of the TRiP(cast) risk assessment tool,¹⁴ so that they can provide evidence-based and personalised advice for individuals prescribed lower limb immobilisation.