The findings, published Monday in the Journal of Clinical Oncology, point to a need for more aggressive policies and strategies for reducing smoking by a new generation of men and women, in all U.S. states.

"Yes, we are making progress in reducing death rates for lung cancer, but there is really a new epidemic and we have to pay attention to increasing death rates in women," said the study's lead author Dr. Ahmedin Jemal, a researcher at the American Cancer Society in Atlanta, Georgia.

Lung cancer currently accounts for about one in four cancer deaths in the U.S., making it the top cancer killer, according to the Centers for Disease Control and Prevention.

But lung cancer deaths among both men and women have been steadily declining since the 1990s - a trend usually credited to public health campaigns and state policies, like cigarette taxes and smoking bans, designed to encourage people to quit smoking and discourage young people from starting.

Previous research has shown, however, that women younger than 50 - which is to say, those born in 1950 and afterwards - were an exception in the recent decline in lung cancer deaths.

Social trends during the 1960s and 1970s, when these women would have taken up smoking in their teens and early adult years are usually blamed for the lung cancer patterns in this group.

But Jemal's team, with colleagues at the National Cancer Institute, looked at national mortality data to see whether there were any regional patterns as well.

The study is based on data for more than one million U.S. white women aged 35 to 84, who died of lung cancer between 1973 and 2007. The researchers compared 23 states, including 10 in the south and six in the Midwest, as well as California and New York.

Between the 1970s and 2007, the risk of dying from lung cancer was highest among women born in the 1930s. Rates then dropped off among women born during the following decade.

When it came to the baby boomers - post-war babies of the late 1940s and 1950s, the numbers of young women dying of lung cancer rose again - but only in some states.

In Alabama, for example, deaths per 100,000 increased from 6.9 to 10.7 among women aged 40 to 44. In contrast, deaths fell from 6.1 to 2.8 per 100,000 in the same age group in California.

Similar patterns emerged in other southern states and in the Midwest, while California and New York showed continued steady declines.

The study is limited by a lack of information on smoking rates and differences in lung cancer care and treatment, noted Dr. Ping Yang, a lung cancer researcher at Mayo Clinic, Florida, who was not involved in the new work.

Jemal said that weak anti-smoking policies could be the reason that more women are dying in the southern and Midwestern states.

"California was the first state to implement a tobacco control program, the first state to have statewide comprehensive ban on smoking in work places. We can only speculate that the decreasing lung death rate (in California) is due to aggressive tobacco control policies, but we can't confirm that," said Jemal.

Still, some southern states like Virginia, a tobacco-producing state, don't show the increase at all, said Dr. William Blot, who has researched cancer rates across the U.S. at the Vanderbilt-Ingram Cancer Center in Rockville, Maryland, and who was not involved in the study.

The good news, Jemal said, is that women who quit smoking in young- and middle-age can reduce their risk of lung cancer compared to those who continue to smoke.

"I think we all know that smoking is the dominant cause of lung cancer, and that rates of smoking have come down. But a sizable section of the U.S. population still smokes and we have to find a way to get the message out that smoking is hazardous," Blot said.

SOURCE: bit.ly/LXvYsg Journal of Clinical Oncology, online June 25, 2012.

• A Permanent Economic Emergency