Discussion

The primary aim of this study was to assess whether psychiatric in-patients are at increased risk of VTE. Nearly a third of patients were judged to be at increased risk, indicating that they would merit some form of intervention on an acute medical ward, such as increasing patient mobility, compression stockings or sub-cutaneous heparin. Our results support those of Choudry and Job,^{Reference Choudry and Job21} who audited VTE risk assessment on old-age psychiatric wards. They found that all patients assessed during the 2 week audit period had at least one risk factor for VTE, and nearly two-thirds of patients (63%) had at least three. None of the patients audited had received a risk assessment. Our findings are also supported by a large study in France, in which 458 psychiatric in-patients were followed for 90 days,^{Reference Delluc, Montavon, Canceil, Carpentier, Nowak and Mercier6} and a study conducted in Japan (N = 190) where doppler ultrasound scanning was used to ensure that even asymptomatic VTE was recorded.^{Reference Ishida, Katagiri, Uchida, Takeuchi, Sakurai and Watanabe5} Both of these studies found a high incidence of VTE among psychiatric in-patients. However, these studies did not record how many of these patients had received a risk assessment from ward staff before each study had started.

In our sample, the majority of patients were prescribed at least one antipsychotic medication. There is evidence to suggest a link between these drugs and VTE risk. Lacut and colleagues used a case–control study (N = 677) to examine the relationship between use of antipsychotic drugs and VTE.^{Reference Lacut, Le Gal, Couturaud, Cornily, Leroyer and Mottier7} They demonstrated a 3.5-fold increased risk of VTE in patients exposed to antipsychotic agents, particularly phenothiazines, butyrophenones and benzamides. A strong association has been identified between chlorpromazine use and VTE, based on a cross-sectional study of more than 29 000 individuals who received antipsychotic drugs.^{Reference Zornberg and Jick22} The risk was greatest during the early stages of treatment. Clozapine has also been implicated in VTE risk, as it has been associated with a fatal pulmonary embolus rate of more than 27 times that seen in the general population.^{Reference Hagg, Spigset and Soderstrom8,Reference Kortepeter, Chen, Knudsen, Dubitsky, Ahmad and Beitz9} Comparison of such studies with those which have quantified risk for well-established VTE risk factors emphasise the importance of addressing psychiatric risk factors to reduce VTE incidence. For example, risk of VTE for women taking the combined contraceptive pill is almost three times that of non-exposed women,^{Reference Vinogradova, Coupland and Hippisley-Cox23} and the presence of malignant neoplasm has an odds ratio of 4.1 for VTE compared with those without cancer.^{Reference Heit, Silverstein, Mohr, Petterson, O'Fallon and Melton24}

In the current study, only a small minority of patients were exposed to frequent physical restraint. However, this factor should still be given consideration. Case reports have suggested that use of physical restraint is another psychiatric in-patient factor which may influence VTE risk.^{Reference Laursen, Jensen, Bolwig and Olsen12} A significant association between physical restraint and the development of VTE has been demonstrated in patients with psychiatric illnesses.^{Reference Tsuda, Sako, Okamoto, Adachi, Hayakawa and Makino25} One study highlighted the effectiveness of prophylactic heparin in 170 patients with schizophrenia, 82% of whom underwent at least one episode of physical restraint.^{Reference De Hert, Einfinger, Scherpenberg, Wampers and Peuskens26} This was a large study (N = 679), although it had a retrospective design and all in-patients were from the same location, which reduces the generalisability of the results. VTE has also been shown to occur in physically restrained psychiatric patients despite pharmacological prophylaxis, suggesting that a thorough VTE assessment and multi-modal prophylaxis may be required in these patients.^{Reference Ishida, Katagiri, Uchida, Takeuchi, Sakurai and Watanabe5} Mechanisms thought to be related to this association include stasis-induced vein wall injury and exaggerated endothelial tissue factor expression.^{Reference Tsuda, Sako, Okamoto, Adachi, Hayakawa and Makino25}

We found that patients on old age psychiatric wards were at significantly greater risk of thromboembolism than those on general adult wards. The difference in the risk profile of these groups is not surprising given the overlap between thromboembolism risk and general ageing, including reduced mobility, as well as other specific comorbidities such as malignancy.^{Reference Heit, O'Fallon, Petterson, Lohse, Silverstein and Mohr27} In our study, almost 40% of those with dementia were at increased risk. Dementia may lead to VTE as it is associated with reduced mobility, frailty and dehydration.^{Reference Kulmala, Nykanen, Manty and Hartikainen28} Moreover, impaired communication may delay diagnosis of VTE.

In addition to the medical contraindications to mechanical or pharmacological VTE prophylaxis, there may be other barriers to prescribing in mental health settings. Nearly a third (27%) of patients in our study were recorded as being at high risk of self-harm. Clinicians may be wary of prescribing compression stockings as they could be used as ligatures^{Reference Hunt, Windfuhr, Shaw, Appleby and Kapur29} and anti-coagulant medications may increase bleeding in individuals who self-harm by cutting.

A minority of patients had a structured VTE risk assessment completed (4.3%) even when risk assessment proformas were readily available. In one centre, the VTE risk assessment proforma was incorporated into all drug charts. The low rates of proforma completion suggest that staff are not aware of the need to measure VTE risk. A higher proportion of patients at the London site had a VTE risk assessment than those in Cardiff or Edinburgh. A local audit of VTE risk had been completed at the London site previously, which may have resulted in greater awareness among staff. In March 2018, NICE released updated guidelines for VTE.⁴ They included a new recommendation that all mental health in-patients should have a VTE risk assessment on admission.

This study has several strengths. We studied a large sample from multiple centres and used a standardised assessment tool. The results should therefore be generalisable to other in-patient psychiatric populations. There are also some limitations. Some in-patient groups were not included, such as CAMHS and addictions units. The VTE risk assessment tool, while recommended by NICE, has not been formally validated in this population. However, we adapted it to include factors relevant to mental health. Assessors were required to judge whether each patient was at an increased risk of VTE based on the information collected. However, they had all undertaken additional training on VTE risk assessment and management, and were supervised by a doctor in psychiatry training. It would have been useful to compare the VTE risk in our patient population with that of general medical or surgical patients. However, audit registration in acute trusts without mental health provision was not feasible. We were unable to ascertain whether those identified as being at an increased risk of VTE went on to develop the condition.

Overall, our findings suggest that psychiatric in-patients are likely to be at increased risk of VTE. Older adults are most likely to be at risk. Further research is required to assess the risk posed by antipsychotics and physical restraint. Specific guidelines for VTE risk management in psychiatric patients and greater staff awareness of VTE risk are required.