Evidence base for venous thromboembolism prophylaxis in acute general hospitals

Numerous interventional studies have investigated the role of mechanical and pharmacological prophylaxis in reducing the risk of venous thromboembolism among those admitted to an acute general hospital. Studies have principally focused on patients undergoing orthopaedic surgery, non-orthopaedic surgery or no surgery. Interventional studies have demonstrated a reduction in symptomatic deep vein thrombosis among patients undergoing orthopaedic surgery who receive LMWH (relative risk (RR) = 0.50, 95% CI 0.43–0.59 ^{Reference Falck-Ytter, Francis, Johanson, Curley, Dahl and Schulman4} ). The use of LMWH in this group is not associated with a significant increase in major bleeding (RR = 0.81, 95% CI 0.38–1.72 ^{Reference Falck-Ytter, Francis, Johanson, Curley, Dahl and Schulman4} ). Prophylactic LMWH is also associated with a reduction in non-fatal symptomatic venous thromboembolism in patients who are undergoing non-orthopaedic surgery (RR = 0.31, 95% CI 0.12–0.81 ^{Reference Gould, Garcia, Wren, Karanicolas, Arcelus and Heit5} ) and possibly a reduction in symptomatic deep vein thrombosis (RR = 0.47, 95% CI 0.22–1.00 ^{Reference Kahn, Lim, Dunn, Cushman, Dentali and Akl6} ) and pulmonary embolism (odds ratio (OR) = 0.70, 95% CI 0.56–0.87 ^{Reference Lederle, Zylla, Macdonald and Wilt7} ) in patients who are not undergoing surgery. However, this is balanced with a significantly increased risk of major bleeding (OR = 1.28, 1.05–1.56 ^{Reference Lederle, Zylla, Macdonald and Wilt7} ). Furthermore, when considering fatal pulmonary embolism or overall mortality, prophylactic LMWH is not associated with significant benefit in any group. ^{Reference Falck-Ytter, Francis, Johanson, Curley, Dahl and Schulman4–Reference Lederle, Zylla, Macdonald and Wilt7}

Evidence base for venous thromboembolism prophylaxis in mental healthcare settings

In contrast to studies in acute general hospitals, there is relatively little published evidence investigating venous thromboembolism incidence and the role of pharmacological or mechanical prophylaxis in mental healthcare settings. A recent observational study that included systematic venous ultrasonography identified deep vein thrombosis in 10 out of 449 participants (2.2%) following 10 days of admission to a psychiatric hospital. ^{Reference Delluc, Montavon, Canceil, Carpentier, Nowak and Mercier8} A total of 16 out of 458 (3.5%) had experienced an episode of venous thromboembolism by 90 days following admission. Of these, three had a non-fatal pulmonary embolism. The study also showed that venous thromboembolism was more likely in older people (8.6% of those aged over 75 years), which may relate to greater exposure to risk factors such as immobility. Another study based on a review of hospital records revealed 17 confirmed cases of venous thromboembolism among 1495 people (1.1%) admitted to an in-patient mental health service for older people. ^{Reference Van Zyl, Wieczorek and Reilly9} This contrasts with an incidence of 2.4% within 91 days among people undergoing total hip arthroplasty surgery ^{Reference White, Zhou and Romano10} and 1.45 per 1000 person years in the general population. ^{Reference Tsai, Cushman, Rosamond, Heckbert, Polak and Folsom11}

There is growing evidence from observational studies that suggests a possible association between antipsychotic medications and venous thromboembolism, particularly for clozapine and first-generation antipsychotics. ^{Reference Hägg, Jönsson and Spigset12} However, it has been difficult to establish whether this association could be a direct pharmacological consequence of antipsychotics (leading to a prothrombotic state) or if it is mediated through other risk factors that are consequences of antipsychotics, such as obesity or sedation leading to reduced mobility. ^{Reference Masopust, Mal and Vali13} Some studies have also pointed towards physical restraint as a potential risk factor for venous thromboembolism in mental healthcare settings. ^{Reference De Hert, Einfinger, Scherpenberg, Wampers and Peuskens14–Reference Ishida, Katagiri, Uchida, Takeuchi, Sakurai and Watanabe16}

Areas of uncertainty

A recent meta-analysis has led to increasing controversy over the potential benefits of pharmacological or mechanical measures to prevent venous thromboembolism among hospital patients who are not undergoing surgery. ^{Reference Lederle, Zylla, Macdonald and Wilt7} Although a reduction in non-fatal symptomatic venous thromboembolism was seen with LMWH prophylaxis, this is balanced with an increased risk of bleeding and no overall benefit in terms of reduced mortality. Furthermore, the relative benefit of prophylaxis only translates to a modest reduction in absolute risk; for every 1000 in-patients treated with LMWH, only three cases of pulmonary embolism are prevented balanced with four additional cases of major bleeding. ^{Reference Lederle, Zylla, Macdonald and Wilt7}

There is also continued uncertainty about the true incidence of clinically significant venous thromboembolism. ^{Reference Greig, Rochow, Crilly and Mangoni17} Although data from epidemiological modelling suggests that venous thromboembolism is responsible for around 60 000 deaths each year in the UK, ¹ data from postmortem studies suggest a much lower rate of around 5680 per year. ^{Reference Kopcke, Harryman, Benbow, Hay and Chalmers18} Whether pharmacological and mechanical prophylaxis could prevent all deaths from venous thromboembolism is also unclear.

Do people who develop venous thromboembolism always need treatment with anticoagulants?

Some observational studies have employed systematic ultrasound screening to identify asymptomatic as well as symptomatic deep vein thrombosis. Although deep vein thrombosis was identified in 10 out of 449 participants following admission to a psychiatric hospital, seven cases were of distal deep vein thrombosis of which only one case was symptomatic. ^{Reference Delluc, Montavon, Canceil, Carpentier, Nowak and Mercier8} The extent to which asymptomatic deep vein thrombosis predisposes an individual to increased risk of mortality remains uncertain, particularly with respect to asymptomatic distal deep vein thrombosis. ^{Reference Vaitkus, Leizorovicz, Cohen, Turpie, Olsson and Goldhaber19}

The advent of computed tomography pulmonary angiography (CTPA) has led to a substantial increase in the radiological diagnosis of pulmonary embolism. ^{Reference Smith-Bindman, Miglioretti, Johnson, Lee, Feigelson and Flynn20} However, uncertainty is growing over the optimum treatment particularly with respect to whether all those with a radiological diagnosis of pulmonary embolism would benefit from anticoagulation. ^{Reference Wiener, Schwartz and Woloshin21} It is thought that small subsegmental emboli may not necessarily be associated with adverse clinical outcomes and that the risks of bleeding from treatment with anticoagulants may outweigh any benefits within this group. ^{Reference Donato, Khoche, Santora and Wagner22}

Benefits and risks of venous thromboembolism prophylaxis in mental healthcare settings

There are no published interventional studies that have investigated the potential benefits of venous thromboembolism prophylaxis in mental healthcare in-patient settings. Despite this, there is ongoing interest in developing and utilising risk-screening tools to identify individuals at increased risk of venous thromboembolism for prophylaxis. ^{Reference Mal, Masopust, Hosák and Konupcíková23} Furthermore, there is no published evidence that has investigated the potential harms of venous thromboembolism prophylaxis in this setting. Although risks of bleeding have been well characterised for people admitted to acute general hospitals, it is not clear whether the same risks apply elsewhere. In particular, prolonged use of LMWH can predispose to thrombocytopenia leading to an increased risk of bleeding. ^{Reference Wang, Honeycutt, Tapson, Moll, Granger and Ohman24} The mean length of stay in an in-patient mental healthcare setting (adult: 52.1 days, older people: 93.2 days) is substantially greater than that of an acute medical unit (5.5 days). ²⁵ With the exception of those taking clozapine, full blood count monitoring is not routinely performed in the mental healthcare in-patient setting. The extent to which staff in mental healthcare settings are trained to administer prophylaxis and recognise potential adverse complications is also unclear. ^{Reference Van Zyl, Wieczorek and Reilly3} For these reasons, it is possible that the risk of thrombocytopenia from LMWH may be greater for those who receive it for venous thromboembolism prophylaxis in the mental healthcare setting.

Balancing the potential risks of bleeding and the potential benefits of preventing venous thromboembolism with pharmacological prophylaxis is problematic. Cost–utility analysis is a method by which the benefits and risks of an intervention may be balanced with respect to quality of life measures. A study investigating the application of cost–utility analysis to venous thromboembolism found that there was a wide degree of variation in individual estimates of cost–utility of both acute venous thromboembolism and bleeding complications from pharmacological prophylaxis. ^{Reference Hogg, Kimpton, Carrier, Coyle, Forgie and Wells26} However, in the mental healthcare in-patient setting, it is sometimes not possible for patients to weigh up benefits and risks of an intervention because of lack of mental capacity. Furthermore, there is little evidence to estimate the potential benefits and risks of venous thromboembolism prophylaxis among individuals who lack capacity as randomised controlled trials have excluded these individuals. ^{Reference Greig, Rochow, Crilly and Mangoni17}