New model predicts future risk of VTE

Researchers have devised a new algorithm to predict a patient's risk of developing venous thromboembolism (VTE) in the next five years, based on simple clinical variables.

Researchers in the England and Wales conducted a prospective open cohort study of primary care patients in 564 general practices, using data that is routinely collected in practice. Participants were aged 25 to 84 years, had no record of pregnancy in the last year or of any VTE, and were not taking oral anticoagulation. There were 2,314,701 patients in the derivation cohort and 1,240,602 in the validation cohort. The main outcome was incident cases of VTE (either deep vein thrombosis [DVT] or pulmonary embolism [PE]) as recorded in primary care records or cause-of-death records. Cox proportional hazards models were used in the derivation cohort to create risk equations at one and five years from baseline. Researchers examined 21 prediction variables based on established risk factors for VTE, specifically those that are recorded in a patient's record and that patients are likely to know. The study was published online Aug. 16 in BMJ.

The VTE rate was 14.6 per 10,000 person years in the derivation cohort and 14.9 per 10,000 person years in the validation cohort. The predictor variables in the final simplified models for both sexes included: smoking status (smoker or non-smoker, and heavy/moderate/light smoker); history of varicose veins, heart failure and chronic kidney disease, any cancer, chronic obstructive pulmonary disease, inflammatory bowel disease, and hospital admission in the past six months; and current use of antipsychotics. For women, current use of tamoxifen, oral contraceptives and hormone replacement therapy were also included in the final model. Variables that didn't change risk, and weren't included in the models, were: current antiplatelet therapy, asthma, cardiovascular disease, atrial fibrillation, and family history of VTE.

The algorithm, embedded in a clinical risk calculator, could be useful in several clinical situations, such as to identify increased VTE risk on or before hospital admission, or before long flights, the authors said. In such cases, prophylaxis could be considered. The algorithm could also be used when considering whether to prescribe medications, such as oral contraceptives, that might increase VTE risk, as well as to identify high risk groups of patients who might need more testing, monitoring or preventive treatment, the authors noted. They cautioned, however, that the model is meant to identify patients at risk of VTE who might require prophylaxis before a medical procedure or other event, not to diagnose symptomatic patients or estimate changing risk during a hospital episode.

A guideline on the issue, “Venous Thromboembolism Prophylaxis in Hospitalized Patients: A Clinical Practice Guideline from the American College of Physicians,” will appear in an upcoming issue of Annals of Internal Medicine.