Cardiovascular risks have been a concern since combined oral contraceptives (OCs) were first introduced. In the past four years new, mostly reassuring information on the safety of modern, low oestrogen dose OCs has become available. However, in 1995 the new information showed higher venous thromboembolism (VTE) risk for OCs containing desogestrel and gestodene compared with levonorgestrel- or norethindrone-containing OCs. The controversial responses by national authorities, their scientific and public health merits were hotly debated and many considered the differences in risk small and resulted from bias and/or confounding. We discuss these arguments and conclude they lack empirical support or cannot account for the 2-fold increased risk. The risk of ischaemic stroke and myocardial infarction (MI) associated with low oestrogen dose OCs are very small in women without cardiovascular risk factors, while increased risk of haemorrhagic stroke is confined to women >35 years of age. Applying the most recent risks to models of OC-attributable events and deaths, OC-attributable mortality in women <35 years is estimated to be <3 per million users annually, rising to about 10 per million users annually among smokers. In the context of external cause mortality (about 90 per million women of reproductive age annually in the UK) such risks appear small. Over the age of 35 years, OC-attributable mortality is a more important concern, particularly among smokers. In the absence of any appreciable OC-attributable mortality in young healthy women, the additional VTE risk for third compared with second generation OCs should be considered when women choose which OC to use.