When we reach the point where obesity threatens life expectancy more than smoking, we are in serious trouble.

That’s what researchers writing for the New England Journal of Medicine say.

They did some fancy statistical footwork and looked at how many years the average person is likely to gain in life expectancy from declining smoking rates. They then plugged in average years lost from rising obesity rates.

For the first time ever, obesity won and now presents a greater threat to life expectancy in America than smoking does.

Let’s say obesity and smoking rates had held steady. In that case, an average 18 year old would expect to gain about 2.98 years in life expectancy over a 15-year period, largely due to a 20% reduction in smoking over the past 15 years.

But a 48% increase in obesity wiped that out. Instead of an increase of almost 3 extra years of life, the average 18 year old can expect less than a year.

So it’s not that life expectancy is expected to fall. It’s still rising. But it’s rising only a fraction of what it would rise if it weren’t for obesity.

“We know that the effects of obesity are not quite as intense as the effects of smoking, but obesity is more widespread,” said Susan T. Stewart, PhD, of Harvard. “It was a little discouraging to see that obesity was winning.”

There’s been about a 1.4% decrease in smoking rates per year in the 15 years prior to 2005. But trends show that from 1971-2006 there’s been an average 0.5% increase per year in Body Mass Index.

Assuming trends continue, 21% of current smokers are expected to quit by 2020.

Unfortunately, the number of “normal weight” Americans is expected to drop by 35% during the same time frame. By 2020, a whopping 45% of Americans are expected to be clinically obese.

Bottom line: the effects of obesity now exceed the effects of smoking on life expectancy.

The researchers cautioned that these kinds of “population-type forecasts” don’t apply at the individual level to people who quit smoking or lose weight.

Do both, and you can really reduce the likelihood of being a medical statistic