A number of case-control studies published in 1995/1996 have shown an apparent increase in the risk of venous thromboembolism (VTE) associated with the use of third-generation oral contraceptives (OC). However, it was discussed very early on that these studies were subject to a number of biases or residual confounding that would have increased the risk estimates for third-generation OC while lowering those for second-generation preparations. Six new studies or analyses were performed trying to take into account many of the methodological problems that were discussed for the initial studies: Two population-based database analyses in the UK and Germany, a new analysis of the General Practice Registry database (GPRD) in the UK, an analysis of a new database of the Transnational study, a re-analysis of the original Transnational study with a new technique, and a population-based study in Denmark. These studies could not confirm a higher VTE risk in users of third-generation OC compared with those using second-generation OC. Data on the risk of arterial thromboembolism (ischaemic stroke and myocardial infarction) show no such difference between generations of OC, with a statistically significant reduction in the risk of acute myocardial infarction from first- to third-generation preparations in one major study. Some of the investigators concluded that there is very likely no increased risk of arterial thromboembolism associated with the use of low-dose oestrogen OC in young women who are properly screened for cardiovascular risk factors or for such conditions. These findings should be taken into account when interpreting the results of studies on the risk of VTE in women taking combined OC.