This is the second in a series of articles describing the steps necessary to produce a maximally optimal GRAS dossier.
Part 2 – Risk Characterization
Recap
In the first of this series, the prospective product developer us urged to heed the carpenter’s cautionary phrase “measure twice and cut once” or the famous IBM sign “THINK AHEAD” by engaging in some preliminary foundational work before commissioning a dossier and encountering delays. The preliminary steps are described in Figure 1.
Introduction – Risk Characterization
The definition of Risk Characterization is the process of combining information from exposure and dose-response assessments to determine the likelihood of harm from exposure to a substance.
Hazard Determination
Hazard Identification
Originally, Hazard Identification was very crude, consisting only of cramming as much substance as possible into a test animal to see what happened. This effort resulted in the now largely ignored LD50 – the dose required to induce mortality in 50% of the animals treated following a single dose of substance. Later however, the process became more refined and sought to determine what changes could be induced at what dosage[2] level, using the most sensitive test available; perhaps more specifically, what dose caused “harm” (to be consistent with the Supreme Court decision that a food could not be seized because the government failed to demonstrate that harm befell consumers exposed to the food).[3]
Thus, the process of Risk Characterization begins with identifying the particular hazard or harm (if any) associated with the candidate ingredient – defining a dose-response relationship. For food ingredients, the maximum amount the animal can tolerate is not as useful as the maximum amount of substance that will not induce change in the animal; that is, you are looking for a No Adverse Effect Level (NOAEL).
Hazard Characterization (Dose-response)
The Hazard must be characterized empirically. For example, imagine that the first test conducted on a new ingredient was the popular 90-day feeding study designed to deliver to rats at dosage levels of 0 (Control) and 1000, 3000 and 5000 mg/kg body weight. Following termination of the study, at which time the following data was gathered: body weights and food consumption (recorded daily), organ weights, hematologic and clinical chemistry parameters and, histopathology – all according to a standardized protocol and conducted under current Good Laboratory Practice. The data is presented as parameters with a statistically significant difference from control values (Table 1).
Table 1. Results of a theoretical 90-day dietary study in rats | |||
Dose Levels with a Statistically Significant Difference from Control Values | |||
Parameters Examined | 1000 mg/kg bw | 3000 mg/kg bw | 5000 mg/kg bw |
Body weight | ND either sex | ND in M; slight decrease in F | Significant decreases in both sexes |
Food consumption | ND either sex | ND in M; transient decrease in F | Significant decreases in both sexes |
Hematology | ND either sex | Scattered results in both sexes | Significant changes in both sexes |
Clinical chemistry | ND either sex | ND in M; Elevated liver enzymes in F | Elevated liver enzymes in both sexes |
Organ weights | ND either sex | ND in M; Decreases in Liver, kidney, thyroid in F | Significant organ weight decreases in both sexes. |
Histopathology | ND either sex | Some minor (age-related?) changes in liver in M; minor changes in liver, kidney in F | Significant findings in liver, kidney, thyroid and sexual organs in both sexes. |
Bw = body weight; ND = no difference from control values; M = males; F = females |
Reviewing the table above, no effects were seen in either sex at the low dose of 1000 mg/kg and obvious effects at the high dose of 5000 mg/kg. The temptation is to immediately call the 1000 mg/kg the NOAEL, but this is a low dose, especially considering FDA likes a safety factor of 100, meaning we would be limited to an acceptable daily intake (ADI) of 10 mg/kg for humans. However, at the mid-dose of 3000 mg/kg, significant results were seen largely in females and nothing in males. Because at the mid-dose, there were few notable effects in males and scattered results in females the temptation is to ignore the female data as not dose-related and name the NOAEL 3000 mg/kg, which allows us an ADI of 30 mg/kg; arguably, an “iffy” proposition.
Now, supposing we find a one-generation reproduction/teratology study in the scientific literature – this is an old study, not conducted to current standards, but appears to be executed well. The results are reported as follows (Table 2):
Table 2. Results of a theoretical one-generation reproduction study in rats | |||
Dose Levels with a Statistically Significant Difference from Control Values | |||
Parameters Examined | 500 mg/kg bw | 2000 mg/kg bw | 3500 mg/kg bw |
Body weight | ND either sex | ND either sex | Small, but significant decreases in both sexes |
Body weight gain | ND either sex | ND either sex | Significant decreases in bw gain gestation days 8-16 |
Food consumption | ND either sex | ND in either sex | Significant decreases in both sexes |
Number of implantations | ND | ND | Reduced |
Number of offspring | ND | ND | Reduced |
Body weight of live offspring | ND | ND | Reduced |
bw = body weight; ND = no difference from control values |
Despite the reproductive study not having hematology or clinical chemistry values or any histopathology, the body weight, bodyweight gain, and food consumption make it comparable to the 90-day study. A comparison of the data mandates the conclusion that somewhere above the 2000 mg/kg dose and 3000 mg/kg, there is an effect on the animals – it may be a matter of toxicity or palatability but regardless, we now have a supportable, greater, NOAEL of 2000 mg/kg, which when the 100-fold safety factor is applied, we have doubled our Acceptable Daily Intake to 20 mg/kg.
Therefore, we have characterized the hazard (any adverse effect on the animals) and determined the NOAEL to be 2000 mg/kg.
Exposure Determination
Exposure as the result of addition to food
In 21CFR170.30(n), there is a list of 43 food categories to which a substance may be added, with the exception of meat, milk, eggs, fruits and vegetables which are to be sold as fresh. These are very general categories such as “Baked goods and baking mixes”, which would include finished goods (pies, cakes, cookies, bread, rolls, etc.) as well as cake mixes, cookie dough, Bisquick®, pancake mix and other mixes one step away from cooking. The other of the 43 categories are just as general and requires the risk assessor to include the consumption of all baked goods even if the candidate ingredient is only for a rarely eaten cookie such as “lady fingers”. The list of categories was generated in 1972 and is woefully outdated, not only in its breadth because new categories have been devised since then and its depth, because the existing categories have been subdivided into hundreds of different foods.
Because the food categories are outdated and have no quantities associated with the food categories, At Burdock Group, we have applied our proprietary software to a much more definitive dataset, USDA’s What We Eat in America (WWEIA) Continuing Survey of Food Intakes by Individuals 2003-2004. This database incorporates the intake profile (amount and frequency) by individuals and produces an accurate estimated daily intake (EDI) of a substance for individuals consuming the food groups selected for the addition of the substance. Estimates of consumption using serving size, results in a grossly inaccurate EDI.
Exposure as the result of the natural presence of the substance
In the best of all worlds, the amount of substance naturally present in meats or produce is estimated and included in the EDI. For a synthetic substance, not existing in nature, there would be no additional exposure.
Aggregate exposure
The aggregate exposure (the estimated daily intake, EDI, expressed in mg/kg for an adult human, at the 90th percentile of consumption) must be less than the ADI – how much less is left to the Expert Panel or regulator. In our example here, an EDI at the 90 percentile (to accommodate the heavy user) of less than 20 mg/kg, would not be a difficult target to reach.
Risk Characterization
At this point, the risk is characterized; that is, the hazard (a safe dose) is compared to the aggregate exposure, and for food ingredients, must meet the standard of proof of safety of generally recognized as safe (GRAS) under the intended conditions of use. This conclusion (or determination of GRAS) must be made by qualified individuals; these qualified individuals may use the 100-fold safety factor but are not bound by it (although the 100-fold safety factor is required for a food additive substance). If the hazard does not meet the 100-fold safety factor criteria and/or there are other considerations such as a vulnerable target audience (e.g., children), necessitating an additional safety factor (e.g., 200-fold or greater). Other considerations might be if the entire ADI was allocated to only one or two food groups and the propensity for some individuals with a hyperfixation on particular foods or taste experiences that are particular pleasing for the individual. Other, more common issues would include foods containing xylitol or sorbitol that may cause stomach upset or diarrhea. In the circumstances mentioned, the risk assessor may ask for a reduction in allocation of the few food groups, although there are other alternatives for (intractable) risk mitigation.
Conclusion
Determining a safe level of consumption is the first step in determining “hazard” and for food, the safe level would be any endpoint in an animal experiment that results in an effect not seen in control animals – the No Adverse Effect Level (NOAEL). The NOAEL should be divided by a safety factor of 100 to arrive at a safe level of exposure for the consumer (the Acceptable Daily Intake, ADI). The amount of substance added to food and from other sources needs to be calculated to determine the Estimated Daily Intake (EDI), which must be less than or equal to the ADI. If however, there are intractable issues with the substance that cannot be resolved with decreasing the amount in the product, there are other means of making the product safe.
Part 3 – Intractable Risk Mitigation Strate - Coming Soon
[1] Caution - this series of articles is not meant to be instructions for a DIY GRAS dossier; in addition, if read closely, the article describes what must be done, but not how to do it. For the “how to” contact Burdock Group.
[2] People often use dose and dosage interchangeably, but these terms have different meanings. Adding the suffix -age to the base word "dose" implies an overall amount. It can be compared to the difference between the words mile and mileage. (https://www.verywellhealth.com/drug-dose-definition-and-examples-1123989 site accessed 13Oct2024)
[3] US v. Lexington Mill & Elevator Company 232 U.S. 399 (1914))
At Burdock Group, we combine unparalleled expertise with meticulous attention to detail in every GRAS notification we handle. Our team of seasoned professionals brings decades of scientific, regulatory, and industry knowledge to the table, ensuring that each submission meets the highest standards of accuracy and compliance. By staying at the forefront of regulatory developments and tailoring our approach to each client's unique needs, we deliver consistently successful results, earning us the trust and respect of clients and regulators alike.
