The Association Between Atherosclerosis and Venous Thrombosis: Results From the Tromsø Study.

It was in 2003 first reported that atherosclerosis was twice as prevalent in patients with unprovoked venous thrombosis as compared with age and sex matched controls. However, two subsequent population based prospective cohort studies failed to confirm this association. In the latter two study designs, atherosclerosis measurements were only registered at baseline, and it could have taken years before venous thrombotic events occurred. Such a time window may have nullified the outcomes of these two studies. In the Tromsø Study, a population based prospective cohort study from Norway, all participants aged 55–74 years and a random 5–10% sample in the other age groups > 24 years, were invited to ultrasound scanning for atherosclerosis assessment at enrollment in 1994–1995 (total n=6651). Atherosclerosis measurements were repeatedly performed in 2001 and 2007, also after a venous thrombotic event occurred.

1) To test whether the presence of atherosclerosis, based on time dependent measurements, increases the risk of venous thrombosis, while taking confounding or modifying factors such as lipid lowering drugs into close account, and 2) whether venous thrombosis increases the risk of atherosclerosis either directly or through confounding or modifying factors.

Carotid intima media thickness (IMT) and total plaque area (TPA) was assessed by ultrasound examination of the right carotid artery. All venous thrombotic events were identified between date of enrollment to December 31, 2010 (n=303). For the statistical analysis for the study question: “Does the presence of atherosclerosis increase the risk of venous thrombosis?”, we analyzed the Tromsø cohort with an extended Cox model, with IMT or TPA as a time-dependent covariate. This analysis was adjusted for age, sex, body mass index (BMI), smoking and use of lipid lowering drugs. For the statistical analysis of the study question: “Does venous thrombosis increase the risk of atherosclerosis?”, we applied a multiple linear regression model in which we compared participants that had venous thrombosis with participants that did not have venous thrombosis during follow-up.

Crude analyses showed that the risk of venous thrombosis increased across quartiles of IMT, which became more apparent when IMT was entered in the model as a time dependent covariate. As compared with participants in the lowest quartile, hazard ratios were 2.01 (95% CI, 1.21–3.36), 2.21 (95% CI, 1.34–3.64) and 3.11 (95% CI, 1.93–4.99), for increasing quartiles of IMT, respectively. As with IMT, there was a dose response relation between increasing levels of TPA and venous thrombosis risk when TPA was entered in the model as a time dependent covariate. However, the associations diminished after adjustment for age, sex, BMI and smoking. Adjusted hazard ratios were, again, slightly higher when IMT or TPA outcomes were entered in the model as time dependent covariates as compared with the crude models, with relative risks of 1 (reference), 1.20 (95% CI, 0.70–2.06), 1.24 (95% CI, 0,71–2.16) and 1.50 (95% CI, 0.85–2.62) for the increasing IMT quartiles, respectively. Risk estimates were not materially affected by provoked or unprovoked venous thrombotic event type. Further adjustments for use of lipid lowering drugs did not significantly affect the associations between IMT and venous thrombosis, or TPA and venous thrombosis, respectively.

Next, we analyzed if atherosclerosis progression was affected by venous thrombosis onset. We observed a steadily increase of both IMT and TPA levels through time in participants without venous thrombosis (IMT 0.861 mm at baseline and 0.979 mm at second follow-up). However, these increases were comparable to increases in IMT/TPA levels in participants that did develop venous thrombosis, after adjustments for age, sex, BMI, smoking and use of lipid-lowering drugs (IMT mean difference 0.002 mm; 95% CI −0.019 to 0.023 at baseline, and −0.011; 95% CI, −0.056 to 0.033 at second follow-up). Timing of venous thrombotic event onset did also not reveal a relationship with atherosclerosis progression.

The presence of atherosclerosis, based on time dependent IMT or TPA measurements, was not a risk factor for venous thrombosis in this study. Furthermore, patients with a venous thrombotic event did not experience increased atherosclerosis progression over time compared to subjects without venous thrombosis.

No relevant conflicts of interest to declare.