Possible noncausal bases for correlations between low concentrations of ambient particulate matter and daily...

Numerous studies of populations living in areas with good air quality have reported correlations between daily average levels of ambient particulate matter (PM) and daily mortality rates. These associations persist at PM levels below current air quality standards and are difficult to reconcile with the toxicology of PM chemical constituents. The unusual level of lethality per unit PM mass predicted by these associations may result from confounding by unmeasured societal, behavioral, or stress factors. Daily average ambient PM levels may be expected to correlate with societal activity level, because a working population increases PM emissions through increased manufacture, power utilization, construction, demolition, farming, and travel. Also, people's perceived and actual health depend on societal and psychological factors. A stress such as anger strongly increases the risk of death due to heart attack. Societal factors modify mortality as shown by calendar-related changes in mortality that are unrelated to air quality. Cardiovascular and respiratory mortality are correlated to day of the week, end of the month, and to the first week of the year. There is likely a role of such nontoxicologic variables in the PM associations, and without vigorously testing if other variables correlate as well as PM, we may erroneously conclude that reducing already low levels of PM will yield real public health benefits.