[dahart]

Sure, life expectancy is the average age at death for all people, and longevity is the average age at which people die of old age.

Death by car crash, for example, affects life expectancy heavily, but doesn't contribute to longevity at all.

All accidental deaths & infant mortality lower life expectancy but not longevity.

[LifeExpectancy]

I was just a reader so far but I had to register now. As someone who works full-time in this field: - You are right about the definition of life expectancy. - I have never ever heard this definition of longevity. It is sometimes used as a synonym for life expectancy. - No one dies of old age.

[dahart]

My sensitivity to the difference between these terms comes from reading Ray Kurzweil's (and many others') bullshit arguments about how humans will eventually become immortal through technology. He intentionally conflates trends in life expectancy with longevity, as do other charlatans who will sell you a cryogenic chamber or life extending elixir. Having life expectancy increase does not mean that large numbers of humans are going to start living to be older than 100 years. The longevity of humans has been around 80 years and barely changing for thousands of years. The life expectancy of different regions has varied wildly depending on how many people die young.

I'm fairly stunned if you work in this field that you wouldn't have heard of the concept of longevity, or at least maximum lifespan, as something distinct and opposed to life expectancy. It's quite important to rule out accidental causes of death if you want to have any chance of understanding how long humans can possibly live. Being able to make statements about diet and exercise requires being able to factor out all car and sports accidents from the average.

Here's what Wikipedia says: "Longevity, maximum lifespan, and life expectancy are not synonyms. Life expectancy is defined statistically as the mean number of years remaining for an individual or a group of people at a given age. Longevity refers to the characteristics of the relatively long life span of some members of a population. Maximum lifespan is the age at death for the longest-lived individual of a species. Moreover, because life expectancy is an average, a particular person may die many years before or many years after the "expected" survival. The term "maximum life span" has a quite different meaning and is more related to longevity."

https://en.m.wikipedia.org/wiki/Life_expectancy

To be fair, the Longevity entry echoes what you said, that some people use it as a synonym.

My hope was that the NIH would be more careful, because they're scientists working in the field, and they know better than to use vague terms that have easily misunderstood meanings.

> No one dies of old age.

True, but you know exactly what I mean, right? Whatever you want to call it, natural causes? Ideal conditions? What is the correct term for people who grow old and don't die of an accident?

[LifeExpectancy]

Thank you. I agree with many of your points. Of course, I know about the discussion of (maximum) life span and also of life expectancy. And, obviously, the maximum and the expected value (life expectancy) of a random variable is something different.

I also agree with your statement that life expectancy depended heavily on how many people die young. Absolutely. But what is not correct is the statement about the number of centenarians. Please check, for instance, the Human Mortality Database yourself at www.mortality.org to see how quickly the number of them is growing. This does not only have something to do with larger birth cohorts entering those ages but also because of major reductions in mortality among people aged 80-100. And I would be very curious if you were able to provide a scientific reference to your statement that human longevity (do you mean maximum life span?) has been around 80 years for thousands of years. Do you agree with me (if you refer to maximum life span) that this is different now?

And -- as you say correctly -- I also agree with you that it would be advisable not to use vague terms. And in my opinion longevity is a vague term since some people use it for life expectancy others for life span. So it would be better if people use those clearly defined concepts to avoid confusion.

My comment about "no one dies of old age" refers to the fact that a certain cause (ICD 10 code) has to be entered on the death certificate. But I also agree with you here: the precision of this information at very high ages might be problematic due to multimorbidity. Your question concerning the "correct term..." I do not want to claim that I know the correct term. In my experience, people usually differentiate between senescent mortality and non-senescent mortality.

Final remark: It seems you are familiar with James Fries' influential paper from 1980 in NEJM [0] since he is talking about "ideal conditions" and "natural deaths", which is pretty close to what you write in your last sentence. :-)

[0] http://www.nejm.org/doi/full/10.1056/NEJM198007173030304

[dahart]

I'm a bit of a hypocrite demanding clarity from the NIH but not being clear enough myself. :P I'm in full and complete agreement with you that vague terms should be avoided. "Mean lifespan" would be better than both life expectancy and longevity. "Top 10% of lifespan distribution" would be clearer than longevity, and probably more informative than "maximum lifespan".

Maybe I should have said that the number of people living to be older than 130 isn't going up, instead of 100. Yes, there are more centenarians now. Yes, there are more people over 80 now. None of that means that we've increased the maximum possible human lifespan in any way. All it proves is that we've decreased the number of people who died before they could have, right? Better medicine, fewer murders, safer cars, cleaner air, less food poisoning. All these advancements help us "live longer", and yet none of them increase the maximum lifespan. We are getting asymptotically closer to the maximum possible, more and more people are approaching the limit, but there is no evidence yet that the limit is moving or has ever moved, and that's all I want to be clear about.

I'm mostly making an argument to counter people (not you) who are, for whatever agenda, intentionally suggesting that increases in life expectancy are due to increasing maximum lifespan. It's a common tactic, and it's a falsehood. The problem here is that the NIH is doing it a little bit. They did paint a picture of huge variability in life expectancy and then conclude that diet and exercise are the major solution.

"But nearly three-fourths of the variation in longevity is accountable to behavioral and metabolic risk factors, including obesity..."

I would be willing to bet that this statistic is citing longevity as I've described it, and not life expectancy. I don't believe it's true that 3/4 of the variability of life expectancy is due to diet and exercise in combination with genetics. But "behavioral" risk factors is super duper fudgy, so I have no idea. Is dying of road rage or a skydiving accident the kind of behavioral risk factor they're talking about? I don't know, because they quoted a statistic that you could interpret to mean almost anything.

If that's the case as I suspect, then this article has knowingly and intentionally mis-used the terms and left a misleading impression without saying something technically untrue, precisely because the popular lay-person's understanding is that they're synonymous.

Scientific references for longevity being constant... I don't have a definitive source, I've mostly had many long discussions about this with my brother who just finished his PhD in anthropology and told me about longevity being constant. Before that I was under the mistaken impression that quotes you get in school about historic people dying at age 35 meant that nobody lived past 40. Lots and lots of people believe this, and it's not true. Here are a few things I got poking around just now:

http://www.livescience.com/10569-human-lifespans-constant-2-...

http://www.ancient-origins.net/news-evolution-human-origins/...

The table in Wikipedia's article gives some indication too (Life expectancy at older age):

https://en.wikipedia.org/wiki/Life_expectancy#Variation_over...

The stats about huge increases in life expectancy for people over age 10, 15 or 21 are all trending in the direction of people who make it to 20 can expect to live to near 80, plus or minus. Much of the "proof" comes from historical writing, and not from scientific evidence, according to my anthropologist brother. But it doesn't take an average to prove people were often living to be four score years old, it just takes a few samples & stories to know it was somewhat common, right?

[Nomentatus]

"life expectancy is the average age at death for all people, and longevity is the average age at which people die of old age."

Which matters if the accident rate is steeply increasing (only, surely) since it's at a historical low. Is your argument that opiods have accomplished this?

[bonzini]

Isn't that mostly accounted by using the median instead of the average?

[dahart]

I don't know; that's a good question. As long as most people die old then probably yes. But longevity is more closely related to the maximum, and median doesn't give you information about the maximum. And there have been times in the last millennium when the median would have reflected high infant mortality, and not the age of people who live a long life and then die of "natural causes". Anyway, nobody really reports medians. Life expectancy is an average measure.

[bonzini]

> And there have been times in the last millennium when the median would have reflected high infant mortality

Sure. During the peak of the Spanish flu, median probably would have been 18 years old or even worse, given infant mortality was close to 50%.

I found some numbers:

* the US had 200.000 deaths (7.7%) for unintentional injury + suicide + homicide

* the "years of potential life lost" are estimated to be 35 on average for unintentional injury + suicide + homicide, or 7 million man years.

* the life expectancy in the US in 2015 was 78.8 years

Putting things together, the life expectancy for people who don't die from unintentional injury + suicide + homicide is 81.5 (78.8 + 35 * 0.077).

I couldn't find data for median age at death in the US. But in LA county in 2012 the median age at death for men was 74 years and 82 years for women, while life expectancy was 79.0 and 83.7 respectively. So median age at death is actually _lower_ than average.

That's actually not surprising if you think about it. If most people die old, there are many people that live even a few years longer than the average. These easily offset the fewer people who die of violent death when computing the average. The US has 5 million people over 85, which account for several tens of millions of man years of "life gained".

Sources:

* https://www.cdc.gov/injury/wisqars/fatal.html

* http://publichealth.lacounty.gov/dca/data/documents/mortalit...

* http://webcache.googleusercontent.com/search?q=cache:bOUTEt_...