“Increased air pollution cuts victims’ lifespan by a decade, costing billions” blared the headline from Eurekalart last summer. I spent five months of the last 15 in Beijing, with arguably the worst air quality in the world. I call Philadelphia my home, with the 10th worst air pollution in the US. In the past, before good statistics were available, I have been an advocate, board member and even expert witness in support of clean air legislation. Now I dreaded discovering what air pollution might be doing to my long-term health. I procrastinated, and left this project on a back burner for a year. But when I finally chained myself to my desk to research this column, the results were not nearly so bad as my fears.
The above Eurekalert article referred to this research from Denmark, and the summary, it turned out was misleading. The question it appears to be asking is, “if you live in a city with 10μg per m3 of particulate pollution, how much sooner must you expect to die?” But in fact, it addresses a different question: “Assume that air pollution has zero effect on the great majority of people, and that the entire burden of increased mortality comes from a small number of unlucky people. If you are one of those unlucky people, how much is your life cut short because of air pollution?” (Even for this unrealistic assumption, I am not convinced that the author did the calculation correctly.)
For context, the study was based on the concentration of the smallest particulate pollution, particles less than 2.5μm in size, which are thought to do the most damage. A concentration of 10μg/m3 for such particles is a level typical of a large American city on an average day. Philadelphia has many days each year exceeding this level. Beijing air on a summer day has 150μg/m3, and winter days are typically 400-600μg. If my reading of the Danish study is correct, it implies that the average citizen of Beijing loses 500 years of life to air pollution.
Beginning my reading, here are the questions I was curious about:
- How much life is being lost to air pollution in American cities and Chinese cities?
- What pollutants are responsible?
- Is the risk linear with pollution, or is there a threshold?
- Are sources of pollution predominantly local or regional?
- Where are the best and worst places to live?
- What diseases are associated with air pollution?
- What can be done to mitigate health consequences of exposure to air pollution?
- Is it better to exercise in polluted air or not exercise at all?
I came away realizing that some of these questions are difficult to address with field studies and epidemiology, and others have not been addressed, even though they are not so difficult. But generally, I was re-assured that air pollution is not as big a health threat as headlines had led me to fear.
How big is the effect overall?
This study looked at day-to-day variations in death rates in Wuhan, a large, polluted city in China’s heartland. They find that 10% of all deaths are due to respiratory disease, and some large fraction of respiratory deaths are triggered by the day’s SO2 level. (Sulfur dioxide is a significant pollutant in China, but not America, because so much coal is burned in and near cities.) This speaks of the immediate effect only, and corresponds to less than one year of life lost. But this kind of study can tell us nothing about long-term effect. Another study in Eastern China (Jiangsu province) compares across cities, so is potentially sensitive to long-term as well as immediate effects. They find a smaller effect of ozone (O3), corresponding to a few months of lost life.
One of my first discoveries in researching an early ScienceBlog column five years ago was that large differences in mortality correspond to small differences in life expectancy. The deep cause of this counter-intuitive effect is the steep rise in mortality curves, building a wall of death into actuarial tables. This is what Benjamin Gompertz realized two centuries ago, but I was a little late to the party.
These mortality statistics are large enough to detect unambiguously, and a few percent increased mortality (up to 10% in China’s most polluted cities) sounds quite serious. But when these numbers are translated into life expectancy changes, the results are far less alarming. 10% in the worst Chinese cities corresponds to less than 1 year of life expectancy. 1% – 2% typical of American cities corresponds to about a month of life expectancy. Much more difficult to quantify is the extent to which the health effects of air pollution are focused on a subset of people who are particularly sensitive, and who will suffer a seriously early death. This is the question addressed by the headline-grabber I quoted at the top of this column [ref].
The most recent comparison of South and North China (where coal was burned freely for winter heat) is featured in Eurekalert with the sensational headline, Air Pollution Cuts 3 Years in Northern China, but the research article behind it reports 8 months.
My own informal study: Life Expectancy in American Cities
Can we see an effect of pollution on life expectancy in America’s largest cities? I looked up the data, and found a surprisingly large variation in life expectancy. Here is a scatterplot of life expectancy plotted against EPA’s measurement of average morning pollution levels for the smallest particles (PM2.5).
There is a correlation that goes in the expected direction, but not statistically significant, and no clear visible trend. For comparison, look at the plot of life expectancy vs per capita income:
Here there is a statistically significant correlation (p=0.01) and a trend that is visible to the eye. Across 25 cities, 29% of the variance in life expectancy can be explained by wealth alone.
[Source for pollution data]
[Source for life expectancy data]
[Source for income data]
Mechanism of long-term damage
When mice breathe air with particulate pollution, their arterial walls thicken and stiffen, arterial plaques increase, and inflammation rises over a period of months [ref]. Similar effects in humans would be expected to increase risk of heart disease and ischemic stroke. Much of this damage is thought to be reversible after ths source of the pollution is removed [ref].
Joel Schwartz of Harvard School of Public Health has persisted through a long career in creating some of the most solid and credible connections between pollution and its health consequences. This classic study, more than two decades old, uses conservative statistical methods to separate effects of weather from pollution. (Weather is known to be highly correlated with daily mortality, more so than pollution, and pollution, of course, is correlated with daily weather and also with season.) The result is a robust conclusion that TSP of 100 μg/m3 increases risk of death by a factor 1.06. The weakness of this finding is that, since the time of this study, TSP=“total suspended particulates” has gone out of fashion as a measure of pollution. TSP measures large particles more heavily than small, but we now know that the smallest particles are most damaging.
Air quality in America has improved in the last 20 years, and most days, most places are compliant with EPA limits. Nevertheless, a difference in mortality rates can be detected between the good days and the bad. A recent study from Schwartz’s group investigated the question of low-level pollutants. They are able to detect effects from three pollutants: PM2.5, O3 and NO2, and report a total ~1% increase in daily mortality.
This is a large unanswered question, very difficult to pose in an epidemiological study design. It is plausible that high exposure for a short time is more damaging than low exposure for a longer time, but the opposite is possible. It is plausible that the combination of chemical irritants (e.g., O3, SO2, NO2 with micron-size particles is worse than either of the two separately, but we don’t know. A “latency” is often assumed, such that today’s exposure to bad air can produce hidden damage that shows up a decade later to cause disease or death. But it is just as plausible that those who are fortunate to escape disease in the immediate aftermath of pollution exposure suffer no long-term consequences. We do know that hospital admissions and both cardiovascular and pulmonary mortality rise in times of major pollution events. But smaller day-to-day fluctuations in air pollution also produce smaller fluctuations in a city’s mortality and morbidity rates, and these can be correlated in long-term studies.
Is there a threshold, below which low levels of pollutants cause no problem?
Probably not. This study by Schwartz found that 1% or 2% of all deaths in Boston are arguably attributed to particulate and ozone pollution, and Boston air is cleaner than most large American cities, and was within EPA guidelines virtually all during the time of the study (2000-2009). A study across different cities in Eastern China also could find no evidence of a “safe threshold”.
Do filter masks do any good?
These cheap, simple respirator masks are a common sight in Beijing. They are so thin that it is easy to imagine that they can’t be doing much of anything, but apparently this simple measure is quite effective. This study from University of São Paulo was based on metabolic response to pollution, and found the response was reduced to undetectable levels by wearing a mask.
Also common in China are indoor air purifiers that continually circulate air through a HEPA filter. The Berkeley Wellness Letter offers some suggestions and emphasizes limitations. A room air purifier provides less effective protection than a mask.
Can B Vitamins Shield you from Harm?
This study looked at short-term effects of particulate pollution only. These include elevated heart rate, suppressed immune function, certain epigenetic changes (DNA methylation), and a decrease in heart rate variability. (The latter is a somewhat mysterious but apparently robust measure of health that has begun to gain recognition as an indicator in recent years [ref].) By all these measures, a modest course of B vitamin supplementation for several weeks preceding exposure completely prevented the physiological response. On the one hand, it’s a very impressive result; on the other hand, what we care most about is long-term damage to the lungs and CV system, and the short-term protection may or may not correspond to long-term protection.
To Exercise or Not to Exercise?
This study finds that the benefits of walking and cycling outweigh the damage done by breathing more polluted air. The claim is that this is overwhelmingly true in moderately polluted Western cities, and remains true in all but the most polluted cities of the developing world. The methodology of the study looks good to me, although the data on which it is based are uncertain. The study doesn’t address high-intensity exercise, which necessarily involves rapid hyperventilation. It is hard to know if lung damage might be caused at an extra-high rate when the body’s cleansing mechanisms are overwhelmed, as they are in cigarette smoking. People in China tend to exercise less on high-pollution days, but when they live in high-pollution cities, they make the most of it and exercise indoors, or outdoors when the pollution is as good as it’s going to get [ref].
The Bottom Line
Mitigating air pollution is an important environmental project, with health benefits that far outweigh the costs. It is indeed a travesty that our EPA is bowing to pressure from GM and Exxon, decade after decade. Mitigation is well worth pursuing in the US, let alone in developing Asian cities. Nevertheless, even in the worst areas of China and India, the air pollution is a major health problem only for a sensitive segment of the population, and overall robs city-dwellers of less than a year off life expectancy.
Two Personal Notes
- I fasted for five days last month, coordinated to end on the Jewish fast day of Yom Kippur. The last two days I took large doses of quercetin, thinking to purge senescent cells. Fasting is supposed to protect normal cells, while sensitizing senescent cells to toxins. Quercetin is a supplement commonly found in health food stores, a flavonoid extracted found in onions and green tea. It has been identified as a senolytic. Results: Difficult to say with any certainty, but I did feel an ease and speed in swimming after I began re-feeding, and perhaps an easing of chronic stiffness in my low back.
- I have a yoga practice that goes back to 1972 and, I believe, has helped me to retain range of motion. The place I feel loss of suppleness most is my lower spine, and MRIs showed a loss of discs beginning 20 years ago. I take daily aspirin, 325 mg at bedtime, and I think I associate this with an easing of flexibility in the low back. Recently, I’ve noticed that if I substitute naproxen (200 mg) for the aspirin, my low back feels less stiff in the morning. Naproxen is a stronger over-the-counter NSAID than aspirin, more likely to produce side effects in sensitive stomachs; some studies claim to detect long-term heart risks. The best reason to prefer aspirin over naproxen is the long history attesting to the safety of aspirin (for most people).
I intend to try more controlled experiments over the next few weeks to see if my first impressions of naproxen’s benefit hold up.