Laura C. Green, Ph.D., D.A.B.T.
Cambridge Environmental
58 Charles Street
Cambridge, Massachusetts 02141

November 2001

Figure 1. Age-, sex-, and race-adjusted population-based mortality rates for 1980 plotted against mean sulfate air pollution levels for 1980. Data from metropolitan areas that correspond approximately to areas used in prospective cohort analysis.

Figure from: Pope, C. A. Ill.; Thun, M. J.; Namboodiri, M. M.; Dochery, D. W.; Evans, J. S.; Speizer, F. E., and Heath, C. W. Jr. (1995). Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. American Journal of Respiratory and Critical Care Medicine. 151:669-674.

Figure 2. Age-, sex-, and race-adjusted population-based mortality rates for 1980 plotted against mean fine particulate air pollution levels for 1979 to 1983. 'Data from metropolitan areas that correspond approximately to areas used in prospective cohort analysis. . . .

Figure from: Pope, C. A. III.; Thun, M. J.; Namboodiri, M. M.; Dochery, D. W.; Evans, J. S.; Speizer, F. E., and Heath, C. W. Jr. (1995). Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. American Journal of Respiratory and Critical Care Medicine. 151:669-674.

Associations between PM ambient air pollution in metropolitan areas and rates of mortality/morbidity in those areas have been noted for some time:

1300's: Edward I and II (persons fouling air with coal smoke to be tortured);

1930's: Firket

1950's: Logan, Stocks;

1960's: Ciocco and Thompson (Clean Air Act of 1963; Federal funding for air quality research; Air Quality Control Act of 1967)

1970's: Lave and Seskin; (Clean Air Act of 1970: National Ambient Air Quality Standards [NAAQS]; globalization; problem solved?)

1980's: Lipfert, Spengler, Ware, Wilson, Wyzga

1990's: Abbey, Burnett, Dockery, Ito, Ostro, Özkaynak, Pope, Roemer, Samet, Schwartz, Thurston; et al.


15 g/m3, annual average standard

(average of three years of quarterly means of 24-hour measurements).

65 g/m3, 24-hour standard

(98th percentile of 24-hour measurements)

NYPA's small, gas-fired power plants

The Supreme Court of the State of New York determined that the New York Power Authority was remiss in not "addressing the health impacts of PM2.5 emissions" of 11 small (< 44 MW) natural gas-powered turbine electric generator units. (July 2001)

The Court noted that NYPA had quantified "PM2.5 emissions by assuming that all PM10 emissions were PM2.5 emissions and concluded that the individual and cumulative impacts of such emissions by the purposed facilities would be insignificant [e.g. max. impacts < 2% of PM2.5 std]." But the Court found that such an analysis was inadequate.

The Court wrote:

"Particulate matter is a nonthreshold pollutant."


"In light of the undisputed potential adverse health effects that can result from PM2.5 emissions, we conclude that NYPA failed to take the requisite 'hard look' at this area of environmental concern."

But, are current levels of ambient PM in fact toxic?

Does this question even make sense?

How can it, since PM2.5 refers to hundreds of thousands of different things?

Differences in:



Chemical composition/pH

Biologic/Immunologic properties

Jamie Robins (2001) writes:

I believe that, in an observational study, every two variables have an unmeasured common cause, and thus there is always some uncontrolled confounding. . . . As epidemiologists, we should always seek highly skeptical subject-matter experts to elaborate the alternative causal theories needed to help keep us from being fooled by noncausal associations.

From: "Data, design, and background knowledge in etiologic inference." Epidemiology 2001 May;12(3):313-20

3 Possibilities:

Deaths associated with ambient PM2.5 are:

1. Caused entirely by PM : Easily controlled

2. Partially caused by PM

Partially confounded by other causes >>>>>>> Pollution

3. Entirely confounded

How toxic is PM/Sulfate according to the observational studies?

Mass-based "toxic potency" values can be derived from the epidemiologic observations. For example, equivalent lifetime Unit Risk Factors (URF's) can be derived from the Relative Risks found in the ACS cohort study re-analysis by Krewksi et al. (2001) 1:

A relative risk of lung cancer of 1.33 (95% c.i. = 1.10 - 1.61) associated with a change in annual average concentrations of sulfate of 19.9 g/m3 (Commentary Table 2) translates to a URF of 1.7 10-2 m3/g.

Such a URF would mean that sulfate in ambient air is:

4x more potent a lung carcinogen than arsenic (URF =4.3 x 10-3 m3/g), and

27 times more potent a lung carcinogen than coke oven emissions (URF = 6.2 x 10-4 m3/g).

How plausible is this?

1. Krewski, D., et al., 2001, Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of Particulate Air Pollution and Mortality. Health Effects Institute; Cambridge, MA

Secondary PM from gas-to-aerosol conversion of NOx and SO2 emissions

  • Visibility concerns
  • Health concerns?
    • sulfuric acid-layered ZnO / ozone (Mary Amdur)
    • "ordinary" sulfates?

Ammonium sulfate and ammonium nitrate are water-soluble
Virtually all metal nitrates soluble
Most metal sulfates also soluble (except Ca, Pb, Hg, few other salts)

Amdur (1986):
" . . . an air quality standard based on "suspended sulfate" without further characterization would be entirely inappropriate; the term is toxicologically meaningless."

Inhaled SO42-: Is it toxic at moderate levels?

  1. Most bronchodilator medications used to treat asthma (such as albuterol, metaproterenol, and terbutaline), are supplied as the sulfate salts.

  2. Each puff from a standard inhaler supplies a metered 100 g of albuterol sulfate, supplying some 20 g of sulfate.

  3. Assume a person breathes in 2 L of air along with each puff of an inhaler.

  4. Thus, 20 g of sulfate per 2 L of air = 10 g of sulfate per L of air, = 10,000 g of sulfate per m3 of air.

  5. Is there evidence that this concentration causes harm?

  6. Mildly acidic sulfate salts have been tested in several systems; inhaled concentrations below several hundred micrograms per cubic meter are NOAELs.

Inhaled SO42-: Is it carcinogenic at moderate levels?

Chronic Cancer Bioassays of Inorganic Sulfate Salts

Sulfate salt vehicle test species result
Aluminum potassium sulfate water mouse


Beryllium sulfate water mouse


Sodium sulfate food mouse negative
Vanadyl sulfate water mouse negative
Zirconium (IV) sulfate water mouse negative

Ambient PM appears to be vastly more toxic, on a mass concentration basis, than cigarette-smoke-derived PM. (Gamble). Is this plausible?

Case-crossover studies

1. Drew Levy, Lianne Sheppard, et al. (2000)

Hypothesis: Risk of cardiac arrest = f [ambient PM]

Subjects: 362 cases of cardiac arrest in Seattle, 1988-94

Air pollution data: nephelometry, PM2.5, PM10, SO2, CO, O3, temp.

Results: Nonpositive. Point estimates suggested that as ambient concentrations of PM increased, relative risk of cardiac arrest decreased.

For an increase of 19.3 ug PM10/m3,

relative risk of cardiac arrest = 0.868

(95% c.i. = 0.744 - 1.012)

Case-crossover studies

2. Peters, Dockery, et al. (2001)

Hypothesis: Risk of myocardial infarction = f [ambient PM]

Subjects: 772 cases of m.i. in Boston, 1995-96

Air pollution data: PM2.5, carbon black, SO2, CO, O3,

Results: Positive.

For an increase of 25 ug PM2.5/m3 2 hrs. prior,

relative risk of m.i. = 1.48

(95% c.i. =1.09 - 2.02)

For an increase of 20 ug PM2.5/m3 24 hrs. prior,

relative risk of m.i. = 1.69

(95% c.i. = 1.13 - 2.34)

But these case crossover studies have not measured behavioral/emotional factors known to precipitate/increase risk of m.i.

Consider a case-crossover analysis performed as part of the Stockholm Heart Epidemiology Program (SHEEP)

Möller et al., (1999)

Hypothesis: Triggering of myocardial infarction

= f [anger]

Subjects: 699 cases of m.i. in Stockholm County, 1993-4

Data on "hostile behavior" and symptoms in days/hours prior to m.i.: gathered through detailed, structured interviews (interviewers blind to hypotheses)

Results: Strongly positive.

During 1 hour after an episode of anger,

relative risk of m.i. = 15.7

(95% c.i. = 7.6 - 32.4)

Thus, if daily/hourly fluctuations in traffic/other activities that increase ambient PM also increase anger, the latter could confound associations between the former and m.i.

1980's: Early hypotheses regarding AIDS/KS:

NO2- to blame?

Hypothesis: Nitrites ("poppers") cause AIDS/AIDS-related Kaposi's sarcoma.

Toxicologic plausibility:

  1. Amyl, butyl, and isobutyl nitrites are toxic at near-typical doses;
  2. Human lymphocytes are damaged (in vitro and in vivo) by volatile nitrites.
  3. N-nitroso compounds are potent carcinogens in all species tested; and
  4. Mutagenic, teratogenic, and carcinogenic products result from the metabolism of N-nitroso compounds;

  5. Epidemiologic plausibility:

    1. Production and sales of volatile nitrites for use as recreational drugs preceded the outbreak of the AIDS/KS epidemic by 7 to 10 years;
    2. Male homosexuals used nitrites extensively; and
    3. Virtually all young male homosexuals with Kaposi's sarcoma had in fact used nitrites.

    4. Far better explanations now thought/known to be:
          HIV; human herpesvirus 8 (HHV8)

Biologic aspects of Particles/Nanoparticles


Sizes: 20 - 250 nanometers

Shapes: two basic - icosahedron (capsid with 20 triangular faces); helix

Constituents: various proteins; nucleic acids

Types: some 3,600 named species in some 164 genera

Properties: benign - virulent

Some Diseases Caused by Viruses
Rabies Tobacco mosaic disease Common cold and flu
Foot and mouth disease in cattle Tomato bushy stunt German measles and mumps
Newcastle disease in chickens Maise dwarf Chickenpox
Distemper in dogs Alfalfa mosaic disease Mononucleosis
Cowpox Sugar beet curly top Cold sores, hepatitis, warts
Influenza in cows, horses, and sheep Dwarfism in rice Herpes and AIDS


Sizes: 50 - 1,000 nm (Some agglomeration in air: most airborne bacteria clumps 8,000 nm)
Shapes: many: cocci, rods, ovoids, spirochetes, filaments
Types: about 5,000 named species in some 800 genera
Properties: benign - virulent

  • Role of microorganisms/infection/inflammation in atherosclerosis/other chronic diseases?

  • Methods for counting airborne microorganisms are inadequate.

Allergens Present in Paved Road Dust in Los Angeles

Allergens (Common Names)
Cladosporium mold
Russian thistle
lambs quarters
mountain cedar
white elm
white pine
white ash
white oak
alternaria mold
meadow fescue grass
dog dander-epithelium
perennial rye grass
western ragweed
Italian cypress
cat dander-epithelium
Bermuda grass
brome grass
house dust mite
natural rubber latex
timothy grass

from: Miguel, A.G., Cass, G.R., Glovsky, M.M., and Weiss, J. (1999). Environ. Sci. Technol. 33:4159-4168.

Chemical compositions of ambient particles in Southern California


(PM0.1 ; averages from samples in seven Southern California cities; from Glen Cass et al., 2000)

PM2.5 (from samples from the San Joaquin Valley; from Desert Research Institute, 1995)
50% organic compounds 36% organic compounds and elemental carbon
8.7% elemental carbon
8.2% sulfate 11% sulfate
6.8% nitrate 34% nitrate
14% trace metal oxides (the most abundant catalytic metals were Fe, Ti, Cr, Zn, and Ce).
3.7% ammonium
0.6% sodium
0.5% chloride
7% "soil"
12% "other"

Ratios of Average Metal Concentrations in aqueous extracts of ambient PM from the Utah Valley:
Steel Mill Operating/Steel Mill Not Operating

Metal Concentration Ratio:

Mill On/Mill Off

Fe 11
Cu 15
Zn 49
Pb 42
Ni 4
V 3

from: Ghio and Devlin, 2001.

Other important differences in metal content in the nonextractable fractions of PM?

Nanoparticles are present in low mass concentrations

(<1 g/m3) but very high numbers (1010 particles m-3)

in urban environments1

Laboratory generated ultrafine particles have been found to produce significant pulmonary inflammatory response in controlled exposure experiments.

However, the study and modeling of nanoPM is complex because:

The effects appear to be dependent on surface properties of the particles,

The effects can be significantly affected by a gaseous co-pollutant such as ozone2,

Measurement of the concentration and composition of these particles is difficult,

Ultrafine or nucleation mode particle are quickly removed from the atmosphere by agglomeration into larger, but still fine, particles, and by deposition,

Ultrafine elemental carbon is a major component of diesel exhaust, and is also an active sorbent.

1. Chung, A., Herner, J.D., Kleeman, M.J., Environ. Sci Technol. 2001; 35(11):2184-2190

2. Oberdorster, G., Int. Arch. Occup. Environ. Health. 2001; 74(1):1-8

What if numbers of particles matter more than their mass?

Then "cleaning up" the air may mean rather different things!

Example (Pitz et al., 2001):

Changes in air quality in 3 communities in East Germany: 1993 vs. 1999

In 1993: several major dominating/polluting industries

By 1999: these industries had closed

1993: PM2.5 mass concentration = 39 ug/m3

1999: PM2.5 mass concentration = 19 ug/m3


number concentration of nanoPM (10 - 30 nm) = 5.9 x 10

number concentration of nanoPM (10 - 30 nm) = 8.2 x 10

Reflects importance of traffic and domestic heating

Is this progress?

Lessons from airborne fiber toxicology:

Numbers, not mass

Biopersistence and dissolution rates important in toxicity

What should we do about ambient air PM2.5?

Stop pretending that PM2.5 is all one thing.

Stop assuming that mass concentrations of all PM2.5 make sense for purposes of regulation/public health improvement.

Think biologically; research accordingly.

Start gathering ambient data on vast number of properties of PM2.5 that might determine its toxicity at ambient levels.

Encourage pollution-oriented epidemiologists and behavioral/pathophysiologically inclined epidemiologists to collaborate. Examine possibilities for confounding more holistically.

Learn from regulatory mistakes: MtBE; ESPs; etc.