Monday, April 6, 2015

Dietary Exposure Assessment

The Basic Equation

As risk assessment issues go, dietary exposure assessment is relatively simple:

Dose = Food Consumption * Concentration

For safety assessments or risk assessments, the concentration is usually known.  So, determining food intake is the only issue.  That is typically done using information from dietary surveys.  The main survey in the United States is:

What We Eat in America (WWEIA), NHANES is a national food survey conducted as a partnership between the U.S. Department of Health and Human Services (DHHS) and the U.S. Department of Agriculture (USDA). WWEIA represents the integration of two nationwide surveys - USDA's Continuing Survey of Food Intakes by Individuals (CSFII) and HHS' NHANES. Under the integrated framework, DHHS is responsible for the sample design and data collection. USDA is responsible for the survey's dietary data collection methodology, development and maintenance of the food and nutrient databases used to code and process the data, and data review and processing.

Extracting the data from the database requires a computer program; there are several available, including one developed by the EPA for estimating pesticide that is available to the public (EPA, 2012).  The resulting output is a statistical distribution that describes the range of intakes in the U.S. population, typically expressed as grams per person or grams per kilogram body weight.

For a safety assessment, one exposure value is chosen to represent the entire population; usually by selecting an upper percentile.  Like the generation of toxicology standards (i.e. ADIs and RfDs), this is a regulatory art that embodies agency policy.  A “Level of Concern” for a particular chemical in a particular food may be calculated as follows:

Level of Concern = ADI / Food Consumption

Like a speed limit, an LOC is a limit intended to reduce risk.  It doesn’t really define a point at which there is a risk (i.e. a threshold), nor does it say anything about how big the risk is.  If you want to know that, then you will need a risk assessment.

Acute Dietary Exposure Assessment

WWEIA is a two day survey. If you want a two day exposure assessment, that is just perfect.  If you want a one day exposure assessment that is no problem either; you can easily generate a distribution of one day exposures. For really short exposures, per serving consumption estimates are easily attainable as well.  While it is not at all common (yeah, regulation can be good), there are many chemicals that occur in food that can have immediate effects.  The short list begins with anticholinergic pesticides that are still used precisely because the only serious human health or environmental concern is from short term exposure. There are also many naturally occurring compounds in food that can have immediate health consequences; cyanogenic glycosides are especially notable.

Since acute toxicity can occur after a single consumption event, characterizing the range of contaminant concentrations is very important.  This can be done with either an empirical distribution or by fitting a statistical distribution.  The former is better when there are many values (>100) and most of the values are above the limit of quantitation (LOQ).  Otherwise a fitted distribution is useful for estimating values both below the LOQ and at higher percentiles.

Chronic Dietary Exposure Assessment

Effects occurring after long term exposure are surely less important; if you are going to be poisoned at all, later is better.  But still, when delving into the realm of what might happen later, WWEIA is less useful.  Since more often than not, food safety issues involve medium to long term effects, that is a problem. 

For estimating lifetime food consumption, the one piece of useful information that can be easily obtained from the WWEIA is average per capita food intake.  Since the survey is designed to characterize the consumption habits of the entire population at all ages, the average intake represents average lifetime exposure for the entire population. For risk assessments where the dose-response relationship is linear, or even approximately linear over the range of dietary exposures, then the average per capita food intake (grams per kg body weight per day) can be used to estimate population risk (e.g. disease frequency). 
There are other surveys that can be used, at least in part, to estimate long term food intake.  In particular:

NCI developed a new instrument called the NHANES Food Frequency Questionnaire (formerly called Food Propensity Questionnaire) and supported its application in the National Health and Nutrition Examination Survey (NHANES). This instrument collects much of the same information as a food frequency questionnaire (FFQ), but without asking about portion size. The resulting data can be employed in the NCI usual dietary intakes model as covariates. By not asking about portion size, the NHANES FFQ requires less respondent burden than other FFQs.

The NHANES FFQ does not over as broad of a range of food categories as WWEIA, but it does cover major staples consumed over a one year period.  Since portion size is not included, the NHANES FFQ must use portion sizes from WWEIA or some other source.  While the portion sizes will not be individually matched, as long as variation in portion size is small relative to the variation that is not serious problem; meaning a reasonable statistical characterization of intakes over the span of a year is possible.  One year is still not the same as “lifetime”, but that is about as good as it gets.

NHANES also includes a detailed 30 day FFQ for fish consumption that characterizes the frequency of consumption of many different species of fish, which is useful for exposure assessments for methylmercury and other contaminants in seafood.

Although estimating food intake is more difficult, characterizing contamination in a food is easier.  Because consuming many different food items over many years will tend to draw from the entire distribution of concentrations that will occur in food, the arithmetic average of the distribution is a good approximation of exposure over a long period of time.

References

U.S. Department of Agriculture (2015).  Agricultural Research Service, Beltsville Human Nutrition Research Center, Food Surveys Research Group (Beltsville, MD) and U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics (Hyattsville, MD). What We Eat in America.

USEPA (2012).  http://www.epa.gov/pesticides/science/deem/.  DEEM-FCID/Calendex Version 4.02/10.00 Release

Official Post Soundtrack

Marley, Bob (1974). Them Belly Full (But We Hungry).  In: Natty Dread, Track 3.

Post Note

Thesis Post #27.  Part of Risk Assessment paradigm thread.

No comments:

Post a Comment