Recommendations for the diagnosis of DVT are based on an evidence-based review produced by the British Committee for Standards in Haematology [Keeling et al, 2004], and are supported by expert opinion published in narrative reviews [Tovey and Wyatt, 2003; Blann and Lip, 2006].

Pregnancy and the postpartum period

• Venous thromboembolism is the main cause of maternal death in the UK, and therefore women who are pregnant or in the puerperium (the period of about 6 weeks following childbirth) require immediate referral if there is any suspicion of risk of DVT [RCOG, 2007].

Intravenous drug users

• Intravenous drug users have usually been excluded from studies investigating the diagnostic accuracy of the Wells Clinical Prediction Rule. However, clinical experience has indicated they are at high risk of DVT, particularly if they have injected in the femoral or groin regions, and require fast-track assessment.

The Wells Clinical Prediction Rule and D-dimer test

• The Wells Clinical Prediction Rule is a clinical tool which was originally developed in 1997 following a cohort study (n = 593) which identified the most important risk factors in people with DVT [Wells et al, 1997].
• CKS recommends the use of a modified version of the Wells Clinical Prediction Rule, which was described in 2003 following a randomized controlled trial (RCT, n = 566) that validated the test [Wells et al, 2003]. The modified test has two advantages compared with the previous test, as it:
• Takes into account previous DVT, which is a major risk factor for subsequent DVT (people who had had a previous DVT were excluded from the original cohort study).
• Has a simplified scoring system, allocating people to one of two groups, 'likely' or 'unlikely'. The original rules allocated people to three groups, with an additional group described as being at moderate or intermediate risk, but in practice management of this group in primary care does not differ from those who are considered to be at high risk.
• There is evidence from several systematic reviews of RCTs that show the Wells Clinical Prediction Rule is an effective method for estimating the risk of DVT. A Health Technology Assessment, Measurement of the clinical and cost-effectiveness of non-invasive diagnostic testing strategies for deep vein thrombosis, used a systematic review and meta-analysis to investigate the role of diagnostic tests (Wells Clinical Probability, plethysmography, D-dimer test, ultrasound, computed tomography, and magnetic resonance imaging) in predicting DVT [Goodacre et al, 2006a].
• Individual clinical features were found to be of limited diagnostic value, but use of the Wells score gave a similar predictive value to that of an unstructured clinical assessment by a physician experienced in the management of DVT.
• Overall the D-dimer test had a sensitivity of 91% and a specificity of 55%. However, the test had better specificity in people who had a low pre-test probability of DVT.
• The authors concluded that 'Further diagnostic testing for patients with a low Wells score and negative D-dimer is unlikely to represent a cost-effective use of resources'.
• A recent RCT has suggested that the Wells Clinical Prediction Rule alone does not adequately rule out DVT in people in the primary care setting [Tagelag and Elley, 2007], which further highlights the importance of the D-dimer test to exclude DVT. Although a cross-sectional study has questioned the value of the Wells Rule in primary care, even when combined with the D-dimer test [Oudega et al, 2005a], methodological limitations of this study means further research in this study population is required [Wells et al, 2006].
• An alternative Primary Care Rule has been specifically designed for the primary care setting [Oudega et al, 2005b; Büller et al, 2009]. However, this has not been directly compared with the Well's Rule, and CKS expert reviewers do not currently recommend it in preference to the Well's Rule.

Referral

• CKS recommends immediate referral or admission for most people where there is a clinical suspicion of deep vein thrombosis (DVT), in line with an evidence-based review produced by the British Committee for Standards in Haematology [Keeling et al, 2004].
• Diagnosis of DVT usually requires confirmation in secondary care by radiological imaging, ultrasonography, and possibly additional D-dimer testing.
• Local referral arrangements vary. For example, referral may be to the local Accident and Emergency department, a medical assessment unit, or a rapid-access vascular assessment unit.
• The D-dimer test is not always available in primary care, and in any case this alone is not usually sufficient to exclude DVT, unless the pre-test probability (that is, absence of clinical features, as measured by the Wells Rule) is considered to be unlikely. For example:
• A high pre-test probability (75%) of DVT and a negative D-dimer test means the likelihood of DVT is about 21%.
• A low pre-test probability (3%) of DVT and a negative D-dimer test means the likelihood of DVT is about 0.3%.

Management in secondary care

• If the person is found to have DVT, they will require treatment in secondary care with anticoagulants and compression stockings.
• Low molecular weight heparin is usually given for at least 5 days, or until warfarin treatment has stabilized the international normalized ratio (INR) to within the normal therapeutic range for 2 successive days [Winter et al, 2005; Baglin et al, 2006].
• There is little evidence from RCTs to support the use of anticoagulation [McManus et al, 2007], with most available trial data having been derived from studies comparing the duration of treatment, rather than placebo-controlled trials. However, there is unequivocal acceptance from experts that anticoagulation is required to prevent complications such as pulmonary embolism.
• In contrast to anticoagulation, there is good evidence from RCTs to support the use of compression stockings, alone or in combination with other treatments, to prevent the recurrence of DVT.

Advice

• Although the likelihood of DVT is very low for a person who has a low clinical prediction score and negative D-dimer test result, it is prudent to advise the person to seek advice if they experience ongoing symptoms, or develop chest and breathing difficulties (since these could indicate pulmonary embolism) [ICSI, 2007].