Facts and Figures

• More than 7 million recognized chemicals are in existence, and approximately 80,000 of them are in common use worldwide (GAO 1994b).

• A 1979 inventory of chemicals mandated by the Toxic Substances Control Act (TSCA) contained 62,000 chemicals that were reported by manufacturers as being in commercial use at that time. The inventory is up to 73,757 chemicals as of February 2001.

• EPA and the FDA have no idea exactly how many chemicals are used in consumer products, nor what products they are used in.

• An unknown number of new chemicals are not among this total. Only new organic chemicals - chemicals that contain carbon - are added to the list. New chemicals that are exempt from the official listing process include inorganics, pesticides, food additives, some large polymer molecules, and any chemical produced in low quantities.

Regulations and Testing Requirements: New vs. Old chemicals

New chemicals

New chemicals are defined here as chemicals that have yet to enter or are still within the Premanufacture Notification (PMN) process. During the 90 day PMN process EPA staff can ask for some limited data on the toxicity and physical characteristics of a chemical, although they rarely do. EPA reviews between 2000 and 2500 applications through the PMN program each year, or between 40 and 50 new chemical applications each week. (EPA 1997, pg. 6 and 44).

• More than half of all PMNs come in the door with no toxicity or environmental fate data. Data on physical and chemical properties like melting point and boiling point may have been submitted for a small percentage of these chemicals (DiCarlo et al 1986, referenced in EPA 1997).

• 8 out of every 10 PMN applications are approved within three weeks, with or without test data (EPA 1997, pg 36).

• 90% of 23,971 PMN chemicals approved by EPA between 1976 and 1994 were approved with no restrictions on their proposed use and production and with no requests for additional test data, regardless of the amount of data submitted.

It is not just health data that is lacking in the PMN process. There is precious little data of any kind submitted with PMNs, chiefly because none is required. As 3M puts it on their PMN forms posted on EPA's web site, "You are not required to submit the listed test data if you do not have it." And chemical manufacturers almost never have it.

• Fewer than 5 of every 100 PMN submissions have any data pertaining to how toxic the chemical might be to wildlife (ecotoxicity data) (Zeeman et al. 1993, referenced in EPA 1997, pg 11).

• Fewer than 4 of every 100 PMN submissions contain any measured values for physical and chemical properties of the chemical (Lynch et al 1991, referenced in EPA 1997, pg 11). This is an extraordinarily significant omission, because EPA routinely uses parameters like a chemical's boiling point and melting point to estimate risks to human health.

• Fewer than 1 of every 100 PMN submissions contain any biodegradation data - an indicator of how persistent the chemical will be in the environment (Boethling and Sabljic 1989, referenced in EPA 1997).

Because it receives virtually no data from industry, EPA relies on estimates of key parameters to judge if the chemical might be toxic to humans, for the vast majority of new chemicals. EPA's standard approach includes a concept called the structure-activity relationship (SAR), which is a comparison of the chemical with other chemicals that are structurally similar for which toxicity data are available. In essence, EPA uses test results for chemical cousins to estimate how toxic a new chemical might be. This is a risky way to do business. The General Accounting Office points to a case where EPA underestimated the risk of a chemical (dialkyldialkoxysilane) by a factor of 100 because they were forced to rely on SARs in their original risk assessment (GAO 1994b).

Old chemicals

Old chemicals are defined here as all chemicals that were on the market when TSCA passed in 1976, plus any chemical that has made it to market through the PMN process. EPA has even less authority to require tests for old chemicals than for new ones. For old chemicals EPA must go through a rulemaking to request a single test on a single chemical.

As of 1998, EPA had requested tests for only 263 chemicals, or only about 0.4 percent of the (approximately) 70,000 "old" chemicals in commercial use in the United States (ED 1998). These are only single tests for one effect, such as cancer, or perhaps just an acute toxicity test. This does not mean that EPA has requested comprehensive testing for 263 chemicals.

What about high production volume chemicals? Isn't industry participating in a wildly successful voluntary test program?

In 1998, EPA reported that the most heavily used chemicals in commerce are largely untested: