A short black a day can keep heart attack at bay, researchers find

JOHN ROSS
November 14, 2017
The Australian
Coffee is more than a heart ­starter, with new research suggesting it helps prevent blood clots in the chest and brain.
American scientists have unearthed fresh evidence that coffee exerts protective effects against heart failure and stroke.
The preliminary findings, presented today at a California conference, support prior research that coffee helps guard against illnesses from cardiovascular disease and cancer to diabetes, depression and dementia.
University of Colorado researchers used machine learning to analyse 70 years of heart health data from the town of Framingham in Massachusetts.
They found regular coffee drinkers proved significantly less likely than ­abstainers to have experienced heart failure or stroke. Each ­additional weekly cuppa, on ­average, reduced the likelihood of heart failure by 7 per cent and stroke by 8 per cent. The team stressed its study had demonstrated only a relationship, not proven cause and effect.
“We don’t know if it is the coffee, compounds in the coffee or behaviour associated with drinking coffee,” said lead author Laura Stevens, a PhD student at Colorado and data scientist with the American Heart Association. “We are still investigating the mechanism behind the ­association.”
Researchers uncovered the relationship by scouring data from the Framingham Heart Study, which has tracked three generations since 1948. The team double-checked findings using traditional analyses of two other data sets, with the association proving consistent in all three studies.
Ms Stevens said the findings demonstrated the potential of machine learning to find new cardiovascular risk factors. “The tools we currently use are very good but not 100 per cent accurate,” she said.
She likened the task to finding five needles in a field of 100 haystacks. The needles were the risk factors, while the haystacks represented broader contexts such as genetics, diet, lifestyle and medical history.
“The ­machine learning algorithm would rank how probable each haystack is to have needles, so you could start by looking in the most likely haystacks,” she said.