Alcohol is killing Americans at a rate not seen in at least 35 years, according to new federal data. Last year, more than 30,000 Americans died from alcohol-induced causes, including alcohol poisoning and cirrhosis, which is primarily caused by alcohol use.

In 2014, there were 9.6 deaths from these alcohol-induced causes per 100,000 people, up 37 percent since 2002.

This tally of alcohol-induced fatalities excludes deaths from drunken driving, other accidents, and homicides committed under the influence of alcohol. If those were included, the annual toll of deaths directly or indirectly caused by alcohol would be closer to 90,000, according to the Centers for Disease Control and Prevention.

In recent years, public health experts have focused extensively on overdose deaths from heroin and prescription painkillers, which have risen rapidly since the early 2000s.

But in 2014, more people died from alcohol-induced causes (30,722) than from overdoses of prescription painkillers and heroin combined (28,647), according to the CDC - although total drug deaths are far higher.

Patterns vary widely from state to state, for reasons that are not always obvious. Alcohol-related death rates in the Philadelphia region are far below the national average - New Jersey's (6.2 per 100,000 residents) was third-lowest in the United States last year, while Pennsylvania's (6.5) was sixth-lowest.

Yet the same set of federal statistics show that Pennsylvania had the sixth-highest rate of drug overdose deaths (21.4 per 100,000 and continuing to rise). New Jersey's drug-death rate (14 per 100,000, just below the national average) was down slightly in 2014 after several years of rapid increases.

Drug fatality rates are even higher in most of the region's suburban counties in both states. But for alcohol, they are lower. In the city, the relationship is flipped.

Philip J. Cook, a Duke University professor who studies alcohol consumption patterns and their effects, notes that per-capita alcohol consumption has been increasing nationally since the late 1990s.

"Since the prevalence of heavy drinking tends to follow closely with per capita consumption, it is likely that one explanation for the growth in alcohol-related deaths is that more people are drinking more," he wrote in an email.

The number of American adults who drink at least monthly rose by a small but significant amount between 2002 and 2014 - from 54.9 percent to 56.9 percent - according to data from the Substance Abuse and Mental Health Services Administration.

The change has been especially pronounced among women. The percent of women drinking monthly or more rose from 47.9 in 2002 to 51.9 in 2014. And the percentage of women reporting binge drinking - defined as five or more drinks on at least one occasion - rose from 15.7 to 17.4 percent over the same period.

Cook notes that when you adjust the alcohol fatality rates for age, the increase narrows somewhat. That's because older Americans are at more risk for alcohol-induced diseases, like cirrhosis, and the U.S. population has gotten older in recent decades.

The heaviest drinkers are at the greatest risk for alcohol-induced causes of mortality. And some drinkers consume plenty of alcohol indeed. Prior research by Cook indicates that the top 10 percent of American adults consume the lion's share of alcohol in this country - close to 74 drinks a week on average.

For people who drink less, alcohol's effects on health are less clear-cut. A large body of research seems to indicate that moderate alcohol consumption - around a drink or two a day - is associated with decreased risk of mortality. But with alcohol, the line between "moderate" and "dangerous" use can be a thin one.

Some researchers want health officials to focus more on the dangers posed by alcohol, and less on the dangers of less toxic drugs, such as marijuana and LSD.

Staff writer Don Sapatkin contributed to this article.