Feb 082012


Just what exactly is the risk of  death from heart and lung disease due to air pollution in Beijing — or anywhere? It’s such an obviously important question that I’m always discouraged and annoyed that I can never find clear answers from any public health group or website. My expat patients always want to know, “with my __  years in Beijing, what’s the long term risk for me and my kids?” And for those who cannot choose to leave easily, i.e. the millions of local Beijingers who are born and raised here, what’s their lifetime risk? You think it’d be an easy question, yes?

No, it’s not. It’s incredibly complicated data to tease out from the research literature. There actually is a lot of data out there, but their conclusions are usually so arcane and user-unfriendly that I find them barely useful. So I’ve tried to make my favorite data more user-friendly, or at least create a starting point for discussion. And I think we do need a robust discussion, as I was actually surprised that the risk was lower than I had expected — at least when compared to other cities. Let’s jump in…

Lies, Damn Lies and Statistics

There are many ways to report mortality data, and perhaps the most user-friendly is life-expectancy data. One of the most famous papers analyzing US data, published in NEJM in 2009 by Dr C Arden Pope, stated that every decrease in PM2.5 of 10ug/m3 increased life expectancy by 0.61 years. Let’s just say this is generally accurate, and that the opposite is true, that an increase in similar PM would cause a decrease in life expectancy; then in Beijing, the average PM2.5 of 100 ug/m3 is about 9 times higher than a WHO recommended annual average of 10: thus 0.6 x 9 = 5.4  years of lost life from long-term living in Beijing. This “five years” number was also mentioned in Steve Andrew’s excellent analysis of the US Embassy data and has frequently been quoted.

But I’m still not happy with that “conclusion”, as this type of analysis is extremely risky:

  • It assumes the dose response is linear and unending into infinity when a lot of data shows it is not, especially for heart disease. Studies implied there may be a leveling off of risk at PM2.5 concentrations over 50 ug/m3, and most of these studies were done at levels under 30, which is far below Beijing’s average.
  • It assumes pollution subtypes are the same everywhere, which is likely untrue. Different molecules carry different risks, especially ones that directly cause cancers. So Beijing’s risks may be different than LA’s risks. In fact, a recent study from Xian showed only 0.2% increase in mortality per 10 ug/m3 increase of PM2.5, which is far lower than other large studies which were done mostly in Western cities much less polluted cities than Xian. These other studies usually showed 0.6-1% increase. If this data is considered accurate and is replicated in other Chinese studies, then is China’s air actually safer than we think (at equal levels of PM2.5)?
  • And just from my gut instinct, that would imply that the average lifespan of all the world’s most super-polluted cities would be clearly at least 5 years shorter than in cleaner cities. And I just haven’t seen this data, nor have I heard of anyone publishing such alarming stats. You think that’d be front page news, no?

There’s another way of looking at mortality data which I think is important to grasp, and that’s via relative risks. It means that you’d have an x% increased risk of developing  a disease when compared to normal risk and disease rate. I really like relative risks (RR for short) also because it’s easy to compare to other diseases and risks. I’ll demonstrate in a minute, but I found my favorite data in another of C Arden Pope’s research here, published just last November. He compared death rates from PM2.5 both from air pollution and also from smoking (secondhand and direct), and the results are important and fascinating.

Lung Cancer Rates Are Not The Same As Heart Disease Rates

The crucial point in Pope’s newer research is that the long-term mortality risks from inhaling PM2.5 are not the same for lung cancer and heart disease:

  • Lung cancer risk increases linearly and dramatically with increasing PM2.5, climbing to a risk of 12.2 for a pack-a-day smoker and 19.8 for a 2-pack-a-day smoker. The equation: [RR = 1+0.3195(dose)0.7433]
  • Heart disease risk increases not linearly but in a parabolic curve, and much less dramatically, with a steep initial curve that tails off and maxes out around 2.5 relative risk. This equation: [RR = 1+ 0.2685(dose)0.2730]

Here below is Pope’s important graph showing long-term mortality risks of PM2.5 on lung cancers (left) and cardiac disease deaths (right). The above formulas are on page 1619 of Pope’s paper.

mortality air pollution smoking C Arden Pope myhealthbeijing

Mortality from air pollution and smoking .C Arden Pope article


At Last: Relative Risks For Air Pollution

Let’s use Pope’s formulas above to calculate what everyone wants to know: the long-term relative risks from different levels of air pollution. I got Beijing’s average AQI from a cool new website, foggybeijing.com, which analyzes the US embassy data from 2010 to present. The other calculations are described in my recent article comparing Beijing’s air to cigarettes:

  • Beijing’s 2011 average PM2.5 AQI of 174 = daily 1.8 mg of inhaled PM2.5 =  lung cancer deaths 1.49 RR (relative risk) and  heart disease deaths 1.32 RR
  • AQI200 = PM 150 = 2.7 mg = lung cancer death 1.67 RR, heart disease death 1.35 RR
  • AQI300 = PM 250 = 4.5 mg = lung cancer death 1.98 RR, heart disease death 1.40 RR
  • AQI500 = PM 500 = 9.0 mg = lung cancer death 2.64 RR, heart disease death 1.49 RR

In other words, living long-term in Beijing (at least 6 years in these studies) causes a 49% increase in lung cancer and 32% increase in heart disease deaths, when compared to perfectly clean air of 0 (although background baseline is 3-5 ug/m3). And even if the average was crazy-bad 500, which doesn’t happen anywhere in the world, the increased risk is around 50-150% higher.

So What Do Those Numbers Really Mean To You?

Is this data high or low for you? Personally for me, it’s lower than I had feared, but let’s break this data down even more, as I think keeping pollution risks in perspective is very important. For example, what’s the risk in Beijing when compared to other historically smoggy cities such as Los Angeles or London? Or in my previous paradise in San Francisco’s wine country? Let’s get the latest annual PM2.5 rankings from the Excel spreadsheet data from the World Health Organization (more info: http://t.cn/aexyqM):

  • LA: 14.8 = 0.27 mg = lung cancer RR 1.12, heart disease death RR 1.19
  • London: 13.5 = 0.24 mg =  lung cancer RR 1.11, heart disease death RR 1.19
  • Paris: 22.9 = 0.41 mg =  lung cancer RR  1.17, heart disease death RR 1.21
  • San Francisco: 9.2 = 0.17mg =  lung cancer RR  1.09, heart disease death RR 1.17

So even in almost-perfect San Francisco (where fog truly is called fog and not smog), the death rate from the almost non-existent air pollution is still 9-17% higher than zero pollution. And in LA, the risks are 12-19% higher; in Paris the risk is around 20%. So for those of you who are constantly thinking of fleeing Beijing to “better” areas, you need to ask yourself: is your next city’s improved risk clinically and physically relevant? In other words, does it make sense to run away from Beijijng’s risks of 32-49% when your next city’s risks are also in double digits?  Perhaps it does, perhaps not. As I said, it’s all relative; I just don’t picture most Parisians sitting in their cafes filled with angst about their 20% risk and wondering if it’s finally time to flee to Aix-en-Provence.

I think this brings up a big point that other people have pointed out: the individual risk of mortality from air pollution can be argued to be relatively small, but the overall population risk is enormous. In other words, each person’s increased risk is quite small (arguably), but when you multiply that one person by tens of millions, that still means there are hundreds of thousands of premature deaths annually in China caused by air pollution.

My Bottom Line

As I said before, I am a bit surprised that the risk isn’t as high as I had feared. Those numbers (1.32, 1.49) are hard numbers I’ve already internalized and accept. I personally think my overall health benefits in Beijing are far greater than the risks — from emotional state of mind to an amazingly cool job to expat perks, et al. Who can say if my increased optimism and energy from my great lifestyle isn’t extending my life more than air pollution is shortening it?

But I definitely am still convinced — more than ever — that the risk of air pollution is real, and I have no intention of throwing away my home air purifiers and N95 masks. I still worry about raising children here, as the data does show permanent lung damage in children. But I never want my readers to forget that we spend 90% of our time indoors, and we have control over our indoor environment with indoor air purifiers. I’ll bet that proper use of these can get your overall daily exposure down to Parisian ambiance.

So, I plan to ride this crazy-fun China wave for at least a few more years, and with continued common sense and preventive measures, I’ll be just fine. And if I ever get lung cancer some day and guiltily think about my past sins (was it the cigarillos?), I still wouldn’t regret it. I am burning brightly and living deeply, and I’m sharing an epic adventure with my wife, the most important person in my life — and I wouldn’t change a thing.

Coming Soon: More Perspective

Next week or two I will follow up with what I think are other important perspectives: comparing air pollution’s risks to other extremely common “lifestyle risks” such as no exercise; high salt diets; lack of fruits and veggies in your diet; obesity; high cholesterol; and smoking. As you will see, their relative risks for mortality are much higher than Beijing air’s 1.32 and 1.49. Especially with smoking, including victims of secondhand smoke; as Pope points out in the same study above, people exposed to secondhand smoke have a relative risk of death as high as 1.3 for lung cancer and 1.27 for heart disease. So where’s the national outrage in China over these victims? Where’s the perspective? Stay tuned…

  8 Responses to “Air Pollution And Risk of Death: Not As Bad As You Think?”

  1. Greetings Richard,

    Thanks for sharing your recent article. I am really happy to see that finally someone is taking the time to do this really critical analysis. Even after reading your article a couple times now and playing with excel for awhile I am still confused. I am exhausted at the moment, so my apologies if there is a straightforward answer to this that I am totally missing.

    Main question: How does variation in combined lung cancer RR and heart disease death RR compare to Years of Life Lost (YOLL) and Pope’s 2009 NEJM study?

    For example, Pope’s 2009 NEJM findings were that a decrease in PM2.5 of 10u/gm3 increased life expectancy by 0.61 years.

    Let’s look at two cities you list:
    Paris with a PM2.5 concentration of 22.9 has a lung cancer RR of 1.17 and a heard disease death RR of 1.21
    and, London with a PM2.5 concentration of 13.5 has a lung cancer RR of 1.11 and a heart death disease RR of 1.19.

    So the ~10ug/m3 (22.9-13.6= 9.4) change between Paris and London resulted in a change in lung cancer RR of 0.06 and a heart death disease RR of 0.02.

    Yet, the ranges between 23 and 13 (approximately) are the same that Pope focused on in his study finding an increase in life expectancy of 0.61 years.

    So how does the lung cancer RR variation of 0.06 and the heart death RR variation of 0.02 convert to the 0.61 years? I see those minor variations in RR and it seems insignificant on a personal level, but 0.61 years is a lot.

    Thanks again for your work, and I look forward to your response.


  2. Hi Steven.

    Great comments! The main issue here is comparing two metrics (life expectancy, or “Years of Life Lost” versus relative risks (RR). I think life expectancy is actually a lot more user-friendly to “consumers” while RR is best for stats experts…Doctor Pope actually has another paper I just got which may finally have some answers: http://ehp03.niehs.nih.gov/article/fetchArticle.action?articleURI=info:doi/10.1289/ehp.00108s4713 If you look at Figure 3 it calculates actual survival curves: so for relative risks of 1.25-1.4, which is what we have above for Beijing, the loss of life is around 2 years.

    Notice how that’s a LOT lower than your “5 year” quote from your excellent paper a few weeks ago in Chinadialogue (http://www.chinadialogue.net/article/show/single/en/4661-Beijing-s-hazardous-blue-sky ) This “5 years” number got a lot of attention in the press but as I was worried about in my article here, the actual data as Pope has here is actually far less dramatic.

    This also brings up another issue; if such answers are already known in a published paper from 10 years ago, why the heck does no reporter or expert ever mention this until now? Why are we ALL scrambling around looking for this data? I think it’s partly a failure of all public health agencies, including the WHO, for not effectively “dumbing down” arcane but important data into “user-friendly” format.

  3. Thanks for your helpful response and posting a link to the Dr. Pope (2000)EHP article. I look forward to reviewing it in more detail. On an initial read, Figure 3, looks quite interesting. I may be reading the text accompanying the figure incorrectly, but it seems to say that a cumulative RR of 1.25 for air pollution reduces life expectancy by 2.5, 2.9, and 3.1 years, respectively, depending on whether exposure starts at 45 years, age 1 year, and at birth, respectively. Where are you getting the “for Beijing, the loss of life is around 2 years“ from in the article?

    Pope concludes that article by saying: “Loss of life estimates due to pollution exposure of 1-3 years for lifelong residents of highly polluted cities, however, is not unreasonable, especially in some of the more polluted cities in the world.” From the WHO data Beijing appears to be one of the most polluted cities in the world, so an estimate of approximately three years of life lost for Beijing would seem to be consistent with the article, including text accompanying Figure 3, and its conclusion.

    “Only” three years of life lost would still work out (roughly) to about a day of life lost per month as a penalty for living in Beijing compared to somewhere with clean air. Does that number sound reasonable?

    My understanding is that in the last dozen years since this Pope (2000) paper was published there has been considerable research finding that the health impacts of air pollution are more significant than previously believe. I need to look into this some more.

    Part of the reason that I used the “five year” number in the China Dialogue article was to be consistent with Zhu Rongji’s comments: “Back in 1999, Chinese Premier Zhu Rongji stated his own fears that air pollution in Beijing would shorten his life “at least five years” – and fine particulate concentrations have not improved since then.”

  4. Steven, you keep me on my toes! Yes, my “2 years” was a bit too casual. This was from a too-casual interpretation of my informal correspondence with Dr Pope, but his actual message was what you said above: about 2.5-3 years of lost life (depending on age) when the relative risk is 1.25. Remember how Beijing’s RR were 1.32 and 1.49, a bit higher, so it’s safe to say the years lost are slightly higher than that — but by exactly how much I’m not sure from the study, as I don’t see the formula in his paper where we can play with the numbers.

    But as for quoting previous Chinese premiers, it may look good in print but did he base that quote on science or was it just a soundbite that became “fact”?

  5. I have been an aerosol analyst with the analytical electron microscope for nearly 40 years. With modern air particles, smaller than even the regulated PM2.5 we face a future like that which we experienced with smoking and asbestos exposure. The problem is that we cannot stop this from happening. It is entirely put of our control.

  6. Thanks for sharing this deep information

  7. Hello Dr. Richard,

    Currently I am looking into graduate schools to attend. Beijing Normal University is one of the schools at the top of my list. I was very excited about the program and prospect of living in China and learning about Chinese culture. But, I been pushed back because of the worsening environmental issues and most peoples negative reviews about life in Beijing. I have never been to Beijing, so I do not have any personal experience to relate peoples opinions too. I am a utilitarian and make my decisions based on analysis.

    That’s where you come in. I would greatly appreciate your personal recommendations about living in Beijing. Do you think that a good educational experience would outweigh the potential health risks associated with living in that environment for 2 years? Any feedback would be greatly appreciated.


    • I would say that especially if only for two years, it’s absolutely a great experience to experience Beijing, in any capacity but certainly as a student as well…

 Leave a Reply



You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>