The serious import of deep venous thrombosis, the effectiveness of treatment and the difficulty of clinical evaluation of this condition make accurate diagnosis essential. Since the specificity of symptoms and signs of deep venous thrombosis is low, most patients who are symptomatic do not have deep venous thrombosis. Anand and associates reviewed the literature and devised a clinical prediction guide to aid in the diagnosis of this condition.

A literature search yielded five studies that addressed the usefulness of clinical findings of deep venous thrombosis compared with venographic confirmation. Risk factors for deep venous thrombosis are recent surgery, immobility, age greater than 60 years, malignancy or cancer chemotherapy, and pregnancy or estrogen use. It was previously thought that careful inspection of the affected leg, looking for tenderness, swelling, redness and Homans' sign, could help diagnose deep venous thrombosis. Physical examination was found not to distinguish patients with deep venous thrombosis from those without it. Impedance plethysmography and compression ultrasonography are two noninvasive techniques used to reliably diagnose proximal but not calf vein thrombosis.

The authors propose estimating the pretest probability of deep venous thrombosis (that is, the probability before impedance plethysmography, compression ultrasonography or venography is performed) using a simplified clinical model (see the accompanying table). The model assigns points based on clinical parameters, such as active cancer, being bedridden for more than three days, pitting edema or an alternative diagnosis that is as likely or more likely than deep venous thrombosis. This pretest probability is then used to determine the diagnostic approach (see the accompanying figure). For example, a patient with a clinical assessment that leads a physician to determine that there is a moderate pretest probability of deep venous thrombosis with an abnormal result on ultrasound examination should begin anticoagulant therapy since deep venous thrombosis is likely.

The authors note that the risk of fatal pulmonary embolus is reduced to less than 1 percent in patients with deep venous thrombosis who are appropriately treated. Application of this prediction guide may help clinicians achieve improved diagnosis and management of deep venous thrombosis.