It is widely assumed by policymakers and health professionals that the harmful health impacts of anthropogenic climate change(1-3) will be partially offset by a decline in excess winter deaths (EWDs) in temperate countries, as winters warm(4-6). Recent UK government reports state that winter warming will decrease EWDs(7,8). Over the past few decades, however, the UK and other temperate countries have simultaneously experienced better housing, improved health care, higher incomes and greater awareness of the risks of cold. The link between winter temperatures and EWDs may therefore no longer be as strong as before. Here we report on the key drivers that underlie year-to-year variations in EWDs. We found that the association of year-to-year variation in EWDs with the number of cold days in winter (<5 degrees C), evident until the mid 1970s, has disappeared, leaving only the incidence of influenza-like illnesses to explain any of the year-to-year variation in EWDs in the past decade. Although EWDs evidently do exist, winter cold severity no longer predicts the numbers affected. We conclude that no evidence exists that EWDs in England and Wales will fall if winters warm with climate change. These findings have important implications for climate change health adaptation policies.