On March 4, 2010, the General Accounting Office (GAO) released a report to the public titled, “FDA Should Strengthen Its Oversight of Food Ingredients Determined to Be Generally Recognized as Safe (GRAS).”[1] The report lamented the fact that there is no requirement in the law or regulation that FDA be informed of a GRAS determination and that the GRAS Notification Process (Federal Register 62:18937-18964, 1997), is inadequate because of its voluntary nature. The GAO’s rationale for compulsory notification was that it felt potentially harmful ingredients could be GRASed and the Food and Drug Administration (FDA) needed to have a better level of understanding, documentation and control of what was being added to the food supply.
Two suggestions were offered for keeping FDA informed, both requiring compulsory notification to FDA of a GRAS determination by either (1) creating a two-tiered system of informal and formal notification or; (2) requiring all GRAS determinations to go through the formalized GRAS Notification Program now in place.
The first, a two-tiered system, would offer two alternatives, both compulsory. Either a substance would go through the GRAS Notification Program as we know it today or, as an alternative, the FDA would be informed of the identity of the substance determined GRAS, but the notifier would not have to submit safety data, nor would there be any input or feedback from FDA – sort of a “notification lite.” Although in its response, FDA rejected the “notification lite” concept, this practice is actually in use today and is exemplified by the arrangement between the Flavor and Extract Manufacturers’ Association (FEMA) and FDA, wherein FEMA only shares information with FDA on the identity and use levels of new flavor ingredients have been declared GRAS. Presumably, there is no feedback from FDA on the FEMA determinations, nor does this information make it to the Center for Food Safety and Applied Nutrition (CFSAN) website.
The second GAO proposed system would consist of requiring all GRAS determinations to go through the GRAS Notification Program, presumably those also performed by FEMA. However, such a requirement would create some serious logistical challenges that CFSAN could not meet without substantive increases in personnel and infrastructure.
As the GAO points out, CFSAN has been relatively responsive to GRAS Notifications within the self-imposed 180 day deadline for the 325 notices posted on the website to date[2], an average of 27 notifications per year. According to the GAO report, from 1998 to 2004 there was a 79% on-time rate (i.e., a response within 180 days), although the on-time rate from 2005 to 2008 fell to 44%. How would this response rate be impacted if all GRAS determinations were subject to compulsory notification?
Conservatively speaking, there are probably at least three to five GRAS determinations conducted for each notified GRAS; therefore, approximately 975 to 1625 GRAS determinations have been made since 1998.[3] Also, FEMA has GRASed 850 ingredients since 1998, with 594 of those ingredients GRASed in the period of 2005 to 2009. Therefore, taking an average over the past 12 years, 152 to 206 GRAS determinations have been made per year (including FEMA determinations). With a recent on-time rate of response at 44% and a submission rate of a minimum of 152 per year, in the first year there would be a backlog of 89 substances, a backlog of 140 the second year and 170 by the third year – FDA would be treading water in no time and drowning in a short time.
This scenario is very reminiscent of the GRAS Affirmation program (initiated in 1977) and it was just such a backlog that killed the affirmation program, resulting in initiation of the voluntary GRAS Notification Process in 1997. In fact, in FDA’s own words:
“In the early 1970s, FDA established a formal GRAS-affirmation petition process in response, in part, to industry requests. This system is now viewed by both industry and FDA as too inefficient and time-consuming. To illustrate: FDA receives an average of 4 to 8 GRAS affirmation petitions per year and acts on an average of 1 to 2 per year. In large part because of their low priority–these substances are nearly always affirmed as safe, and are being marketed while FDA reviews the petition–GRAS petitions can take between 6 and 10 years to review, which has led to a backlog of about 80 GRAS petitions.” Reinventing Food Regulations (January, 1996) http://govinfo.library.unt.edu/npr/library/rsreport/foodreg.html (accessed 21Oct09)
Therefore, if FDA could not deal with as few as 4 to 8 GRAS affirmation petitions per year, how could it be expected to deal with 152 per year? Now while FDA may claim that it does not examine a Notification to the level of detail that was done with an Affirmation, the seductive nature of data contained in a Notification is too much for FDA to ignore. The proof of recent more detailed scrutiny of Notifications is obvious with the slow down in response times as the Notification program proceeded. However, FDA’s response to the presentation of data is understandable in the context of protecting the consumer – that is, if data is presented, CFSAN is obligated to examine it, lest a notifier tries to slip something under the nose of the agency. Can the FDA hold itself to a promise to not look at data? Probably not, because the primary reason for GAO to suggest compulsory notification, is the claim that FDA needs to strengthen its oversight of GRAS determinations. Compulsory Notification is simply not a workable concept.
While the GAO report was honorable in its intentions, the report was not as thoroughly researched as it could have been and, resulted in presenting suggestions that although intuitively obvious, are not workable in the context of practical application.
References
[2] March 15, 2010
[3] Not all GRAS determinations are for new substances, some are just changes in the parameters of use for a previously approved substance, including but not limited to changes in use levels, addition of a food category, addition of a technical effect, a change in specifications, etc.