Background: The role of varicose veins (VV) as a risk factor for development of deep venous thrombosis (DVT) is still controversial. The aim of this study in primary care was to determine the impact of varicosity as a potential risk factor for developing DVT.
Patients and methods: During the observation period between 01-Jan-2008 and 01-Jan-2011, all cases with VV (ICD code I83.9) and DVT (ICD codes I80.1 - I80.9) were identified out of the CONTENT primary care register (Heidelberg, Germany). The exposure of VV and DVT was based solely on ICD coding without regarding the accuracy of the diagnosis. The covariates age, gender, surgery, hospitalization, congestive heart failure, malignancy, pregnancy, hormonal therapy, and respiratory infection were extracted for each patient. Multivariate binary logistic regression was performed in order to assess potential risk factors for DVT. The SAS procedure "PROC GENMOD" (SAS version 9.2, 64-bit) was parameterised accordingly. A potential cluster effect (patients within practices) was regarded in the regression model.
Results: There were 132 out of 2,357 (5.6 %) DVT episodes among patients with VV compared to 728 out of 80,588 (0.9 %) in the patient cohort without VV (p < 0.0001). An increased risk of DVT was associated with previous DVT (adjusted odds ratio (OR): 9.07, 95 % confidence interval (CI): 7.78 - 10.91), VV (OR 7.33 [CI 6.14 - 8.74]), hospitalization during the last 6 months (OR 1.69 [CI 1.29 - 2.22]), malignancy (OR 1.55 [CI 1.19 - 2.02]), and age (OR 1.02 [CI 1.01 - 1.03]).
Conclusions: There are strong associations between VV and DVT in a general practice population with documented VV. Special medical attention is required for patients with VV, a history of previous venous thromboembolism, comorbid malignancy, and recent hospital discharge, particularly those with a combination of these factors.