Quantitation Principles And Methods

Cancer risk assessment has traditionally used the linear multistage model for low-dose risk estimation. However, if a nongenotoxic mode of action is accepted for a chemical, linear low-dose extrapolation methods may not necessarily be appropriate. Risk assessment based on non-cancer endpoints has generally involved determining the lowest observed effect level (LOEL) or no observed effect level (NOEL) and dividing by uncertainty factors (UFs) to estimate a dose likely to protect humans from adverse effects. This extrapolation considers uncertainty and variability, including sensitive population groups. Safe levels can be estimated for acute, subchronic, and chronic exposures by the same principles. Thus, if there is a good study in healthy adult humans that demonstrates a safe chronic dose, dividing that dose by 10 may be adequate to protect sensitive populations (3). Additional factors of 10 for other extrapolations (estimating safe chronic dose from acute or subchronic data, estimating human dose from animal data, accounting for an incomplete database or severity of adverse effects) can account for further uncertainties. The "factor of 10" convention is undoubtedly a very crude estimate of total uncertainty (4-7). This approach also does not calculate the actual risk of an adverse effect; it only provides a rough estimate of maximum safe dose.

To improve this situation, dose-response and cross-species extrapolation models have been developed and applied to risk assessment. Physiologically based pharmacokinetic (PBPK) models (8, 9), benchmark dose methods (10-12), and advanced threshold curve-fitting models for continuous data (13) have all contributed to the development of more quantitatively based risk assessment methods. Although there has been tremendous progress in modeling physiological processes and toxic responses, consensus has by no means been reached on how to estimate safe dose in humans from animal data (14, 15). The debates are even more heated when the endpoint is increased tumors, with a proposed non-genotoxic mode of action.

Risk assessments for nongenotoxic carcinogens may be carried out using any of the above methods, from NOEL with uncertainty factors, through complex threshold data-fitting methods, to linearized multistage. It is common for several different quantitation methods to be explored. The most defensible method, or sometimes the most health-protective conclusion, is chosen.

The methods must follow the conclusions reached in the hazard assessment as to the potential mode(s) of action. The assessment begins by analyzing the data in the range of observation and evaluating how best to extrapolate to low-dose, health-protective risk levels. If standard cancer models are used, the de minimis risk criteria are usually set at the 95% lower confidence limit of dose yielding a 10-6 risk level, as for genotoxic carcinogens (see chapter on genotoxic carcinogens). When UF methods are used, the methods are applied as for noncar-

cinogens except that an additional tenfold safety factor may be applied for severity of effect. When animal data are the basis of analysis, estimation of a human dose may use PBPK modeling if appropriate data are available.

For most toxicity tests, about 10% of the animals responding (the ED10) represents a practical observational threshold, the lower limit for statistical significance. Alternatively, a 10% average change in a measured parameter may be used. The estimated dose corresponding to a 10% response by either definition represents a common "benchmark" from which an effective safe dose may be extrapolated, either with probabilistic methods or by application of uncertainty factors. Depending on the magnitude of the combined uncertainty factors, the effect of the threshold assumption is usually a higher estimated safe exposure level than if traditional low-dose linear extrapolation is used. The difference in dose response extrapolation between nonthreshold and threshold mechanisms is illustrated in Figure 1 below. In this figure, the squares represent percent of tumor-bearing animals in a lifetime cancer study. The solid line represents a fitted curve through zero dose, and the dotted line represents a straight line drawn through the two data points with a significantly increased proportion of tumors.

Using the linear multistage model (the solid line), the upper 95% confidence limit on the dose that produces a 10-6 cancer risk for these data is 0.01 mg/kg. Using a 100-fold uncertainty factor from the threshold of 500 mg/kg, defined by the dotted line, provides an estimated safe dose of 5 mg/kg. Specific corrections for differences among species (such as metabolism, distribution, and

Fig. 1. Theoretical tumor data, illustrating extrapolation of cancer risk using a curve-fitting model (solid line) drawn through zero dose, and a straight line drawn through the points with significantly increased tumor incidence (the dotted line), defining a threshold at a dose of 500 mg/kg.

tissue sensitivity) and experimental uncertainties usually retain this large difference between nonthreshold and threshold estimates of a safe dose.

Quantitative estimates of a safe dose are usually affected more by the decision to apply a non-linear low-dose assumption than by the method chosen among the linear or non-linear models. The examples provided below show how these principles have been applied in some recent risk assessment decisions that involve primarily the U.S. Environmental Protection Agency (EPA) and the State of California.

10 Ways To Fight Off Cancer

10 Ways To Fight Off Cancer

Learning About 10 Ways Fight Off Cancer Can Have Amazing Benefits For Your Life The Best Tips On How To Keep This Killer At Bay Discovering that you or a loved one has cancer can be utterly terrifying. All the same, once you comprehend the causes of cancer and learn how to reverse those causes, you or your loved one may have more than a fighting chance of beating out cancer.

Get My Free Ebook

Post a comment