Those who always seem to see the proverbial glass as half-full are more likely to outlive those who see it as half-empty.

U.S. researchers reported at the American Psychosomatic Society's annual meeting Thursday that they found a strong correlation between optimism and a person's risk for early death, cancer, and heart disease.

The findings came out of the massive Women's Health Initiative study run by the National Institutes of Health, which has followed more than 100,000 women ages 50 and over since 1994.

Researchers at University of Pittsburgh reviewed questionnaires that surveyed the women's personality traits, and then tracked their rates of death and chronic health conditions for an average of eight years.

Women who were found to be optimistic -- those who said they expect good rather than bad things to happen -- were 14 per cent less likely to die from any cause compared to pessimists, and 30 per cent less likely to die from heart disease.

Women who scored high on scales measuring their cynicism and hostility had a higher general death rate and a 23 per cent greater risk of dying from a cancer-related condition by the end of the study.

Cynical and hostile women tended to agree with statements such as: "I've often had to take orders from someone who didn't know as much as I did" or "It's safest to trust nobody."

The research adds to similar findings from previous studies that have linked optimism to longer life.

But Hilary Tindle, the lead author of the study and an assistant professor of medicine at the University of Pittsburgh, says the research team could not tell if optimism itself led to reduced health problems perhaps because of lower stress levels, or whether optimism led to healthier lifestyle - or a combination of both.

"What is the link? What is the mechanism? That's one thing my study can't answer," Tindle said.

Tindle noted that previous research suggests that optimistic women tended to have a healthier "risk profile."

"They are less likely to smoke, they are more likely to be active and they are more likely to have a lower BMI [body mass index]," she said. "All of these are risk factors that certainly matter for length of life and health."

Tindle said her team tried to cancel out the statistical influence that behaviours such as smoking or exercise had on death risk but still came up with the same findings.

"Even after for controlling all of those factors, we found a link," said Tindle.