London - Eating red meat just twice a week raises the risk of early death, according to a major 30-year study.
Processed food such as bacon and sausages and unprocessed steaks and chops were all found to raise mortality rates.
Chicken increased the risk of heart disease, but not the chance of early death. Eating fish had no adverse impact on health.
The US researchers – from Northwestern University in Chicago and Cornell University in New York state – tracked nearly 30 000 people for three decades.
They found eating two servings of processed or unprocessed red meat increased by three percent the risk of dying of any cause of death within the study period.
"It’s a small difference, but it’s worth trying to reduce red meat and processed meat like pepperoni and deli meats," said Norrina Allen of Northwestern.
"Red meat consumption also is consistently linked to other health problems like cancer."
Eating poultry twice a week was linked to an increase in cardiovascular disease of four percent.
But the researchers said this finding was less clear-cut – and may be linked to how someone cooks their chicken. Deep-fried chicken, for example, may be more harmful than roast poultry.
Dr Victor Zhong of Cornell said: "Modifying intake of these animal protein foods may be an important strategy to help reduce the risk of cardiovascular disease and premature death. Our study shows the link to cardiovascular disease and mortality was robust."
Published in the JAMA Internal Medicine journal, the research suggests eating processed meat twice a week increased the risk of cardiovascular disease by seven percent, and by four percent for unprocessed red meat. The participants, who were aged 53 on average at the beginning of the study, were asked about their typical food intake and then tracked for 30 years.
The researchers defined a serving of unprocessed red meat or poultry as a 4oz portion (115g). For processed meat, a serving was two slices of bacon, two small sausages or a hot dog.
Professor Linda Van Horn of Cornell said: "Fish, seafood and plant-based sources of protein such as nuts and legumes, including beans and peas, are excellent alternatives to meat and are under-consumed."
UK researchers welcomed the findings but said the results relied on a self-reported diet, which is notoriously inaccurate.
Professor Gunter Kuhnle of Reading University said: "The increase in absolute risk is so small that it is unlikely to be relevant for the individual.
"On a population level, this is more important. With about one million people being diagnosed with heart disease every year, even a small reduction in absolute risk can have a considerable effect."