The National Health and Nutrition Examination Survey (NHANES) [draft-19APR]

What is NHANES?

The National Health and Nutrition Examination Survey (NHANES) is a program that combines interview and physical examination results from volunteers to assess the health and nutritional status of the U.S. population. The NHANES started in the early 1960s, and has surveyed over 190,000 people. The NHANES conducts interviews of participants to collect information about participant demographics (e.g., age, sex, race, ethnicity, language, etc.), socioeconomics (e.g., education, occupation, income, marital status, etc.), dietary habits, and general health status. The NHANES program collects information about the U.S. population health, dental and physiological status during medical exams. Findings from the NHANES interviews and medical exams are used to help guide decisions and policy around public health disease and risk factors.

One example of the value of NHANES is the use of blood lead data that develop policies to reduce exposure from lead. Blood lead data helped develop policy to eliminate lead from gasoline and in food and soft drink cans, which resulted in a reduction of blood lead levels of more than 70% since the early years of NHANES study.

An important part of the NHANES is the investigation of environmental chemical exposures. In 2009, the Centers for Disease Control and Prevention (CDC) published a report entitled “Fourth National Report on Human Exposure to Environmental Chemicals, 2009” (usually called “the Fourth Report” for short). That report included statistics on exposure status for 212 chemicals. Since 2009, “updated tables” have been published every few years. The most recent set of updated tables were published in 2017 and included the statistical assessment of NHANES data collected during 2013 and 2014. The large collection of data has been divided and published in two volumes. Not all chemicals are measured in every sampling year, and new chemicals have been added since the Fourth Report was published.

The Four Corners States Biomonitoring Consortium (4CSBC) is using the statistical assessment for NHANES 2011-2012 collection period. The 2011-2012 data was the latest available when the 4CSBC started in 2014. In a few cases, the 4CSBC is testing for an exposure that was not measured by NHANES during the 2011-2012 collection period, but was measured in an earlier period. In those cases, we are using the earlier measures.

How is the NHANES conducted?

The actual assessment of participants is conducted by CDC in a collection of mobile trailers. These trailers contain clinics and labs that support the NHANES assessment process. To conduct the survey, CDC moves the trailers from location to location over the course of the two-year sampling period. Only a few locations are selected across the nation because of the complexity of coordinating testing sites, recruiting participants, and conducting surveys.

Selection of the sampling locations is not simple (random). Rather, it involves a complex multi-staged process to ensure that the final participant pool is representative of the national non-institutionalized (i.e., not in the military, nursing homes, prisons, etc.) population. To start, the United States is divided into 15 geographic groups based on their demographic characteristics. From each group a county (or sometimes a cluster of small counties) is selected. Within the selected counties, between 20 and 24 communities or neighborhoods are identified. From those neighborhoods, about 30 households are selected. Finally, selected individuals from each household, who are able to participate, are invited to be part of the NHANES survey population. The average number of participants is about 2,600 persons per sampling period for the metal exposure assessment.

In exchange for spending a few hours visiting the NHANES trailers, participant received a complete physical and dental health exam.

The 95th percentile: What does it mean?

The NHANES data can be thought of as a range of exposures that most people are normally exposed to. For example, all of us have some exposure to arsenic. Since arsenic has no nutritional value and can cause harm to our bodies, our bodies get rid of it. One way to get rid of it is through the urine. As a result all of us have some arsenic in our bodies and in our urine, at all times.

The NHANES data represents the range of results collected from generally healthy people. Statisticians make calculations to create values that help understand the data. For example, the average can be calculated. For arsenic, the average urinary level is around 8 μg/g creatinine (see explanation of this unit of measure in the next section). One way to look at an average is that about 50% of the people tested had urine arsenic levels below 8 μg/g and about 50% of the people were above it. In addition to calculating the average, statisticians can also look at the lowest and highest level. Even though we all have some arsenic in our urine all the time, the lowest level can be so low that the test methods used cannot measure it. Those are reportable as “below the analytical limit of detection” (or “<LOD”). Obviously, one person (or sometimes a few) is going to have the highest level. That person's level will be the maximum level found for the participant population. To protect that person’s identity, statisticians use 95th percentile (95%) level – which is just slightly lower than the maximum level observed, to represent the range of all results collected from all participants. For arsenic, this was found to be 52 μg/g creatinine. Thus, the range usually is presented as the average (or 50th percentile) and the 95th percentile. For arsenic, we could say that 50% of the participants had levels from <LOD to 8 μg/g, and 50% of the participants had levels from 8.1 to 52 μg/g creatinine. About 5% of the population have levels above the 95th percentile. The 95th percentile can be thought of as the "high end of normal."

The NHANES 95th percentile is not exactly consistent through time. Instead, it wanders up and down, depending on a variety of factors. Those factors may include study influences, such as the particular selection of populations from which the participants were drawn from, to national factors such as recent changes in national behavior (i.e., increased or decreased overall seafood consumption due to changes in economics, or the marketing of a new arsenic removing technology). Typically the variability in the 95th percentile is small (50 to 80) rather than large (say, 5 to 500). Thus, one should not think of a 95th percentile of 52 μg/g creatinine as a “thin red line” that separates low normal from high normal. Rather, it providess an idea of where that threshold lies.

NHANES and health implications.

The NHANES 95th percentile has NO clinical meaning. People who have levels above the NHANES 95th percentile level could be perfectly healthy. They might also be experiencing poor health – but for other reasons. In addition, one urine sample is a single “snap-shot” measurement which might have been taken at a moment when that participant’s levels were higher than typical for that participant. For example, urine arsenic levels tend to be higher shortly after eating certain kinds of seafood. One serving of seafood can result in urinary arsenic concentrations that are 200 times higher than what would be typical for a person with just their normal environmental exposures.

Typically, the NHANES 95th percentile level is well below any threshold of concern for health reasons. Clinical assessment for arsenic toxicity typically does not use urine. Blood is a much better sample for testing arsenic for health concerns. This fact holds for other metals and other exposures that the 4CSBC is testing for. Although blood is a better sample for clinical assessment, urine is more convenient for biomonitoring purposes. The biomonitoring results are useful to help guide decisions about policies and programs to reduce exposure. Biomonitoring results should not be used to make a health assessment or for diagnostic purposes.

For example: A person with a 4CSBC urinary arsenic result higher than the NHANES 95th percentile might want to consider how they might have been exposed. The first question to ask oneself is whether the result represents a single exposure event (such as one seafood meal the night before participating in the project) or a continuous exposure (such as daily seafood meals, or a contaminated well, etc.). The factsheets provided with the results letter will help participants through this thought process. If the participant determines that there might be a continuous exposure, the follow-up questions should focus on and explore options to reduce exposure (reduce the number of seafood meals, treat the well water, etc.). The 4CSBC uses the NHANES 95th percentile value as a comparison value to help participants explore these decisions. Visit our page on arsenic for more information about arsenic specifically.

Rarely, a urine sample actually indicates that the participant is experiencing harm from exposure. If the participant is experiencing harm from exposure they should consult with their private health care provider. Their doctor may want to do a second test or do a different (i.e., blood) test to make sure that the elevated status is persistent and get a better measure of how high it really is. Unfortunately, the 4CSBC project is testing for chemicals that do not currently have any reliable indicators of what that level is for urine.

The 4CSBC is conducting biomonitoring on a few chemicals for which there is very little information available about the health effect that they may be associated with. In other words, we don’t know if these chemicals are important causes of health concern and if so, the level at which that chemical may make someone sick. Phthalates are examples of these kinds of chemicals. Trying to quantify the exposure levels so that future studies will have a starting point to work from is one of the true values of biomonitoring. However, for now, we may be unable to tell you very much about what your results mean.

Table of NHANES 95th Pecentiles used by the 4CSBC


all values are in μg/g creatinine

All Persons Only Children (6 to 11 years) Only Adults (20 years & older)
Total Arsenic 50.4 91.2 49.7
Cadmium 0.907 0.235 0.977
Manganese 0.548 0.633 0.545
Mercury 1.75 1.62 1.95
Selenium This analyte has not been tested by NHANES
Uranium 0.029 0.031 0.029
Phthalate Metabolites
Mono-benzyl phthalate (MBzP) 26.7 81.0 21.6
Mono-n-butyl phthalate (MnBP) 41.2 73.3 36.2
Mono-isobutyl phthalate (MiBP) 27.5 46.4 25.2
Mono-ethyl phthalate (MEP) 541 177 634
Mono-methyl phthalate (MMP) 11.0 30.2 9.31
Mono-2-ethylhexyl phthalate (MEHP) 8.47 7.15 8.49
Mono-(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP) 37.7 58.1 32.0
Mono-(2-ethyl-5-oxohexyl) phthalate (MEOHP) 23.4 34.5 19.3
Mono-(2-ethyl-5-carboxypentyl) phthalate (MECPP) 59.8 84.4 53.7
Mono-isononyl phthalate (MiNP) 17.6 7.88 18.1
Mono-(carboxyoctyl) phthalate (MCOP) 194 161 194

Mono-n-octyl phthalate (MOP)

[from the 1999-2000 NHANES period]




Mono-(3-carboxypropyl) phthalate (MCPP) 36.6 28.2 36.2
Cyclohexane-1,2-dicarboxylic acid-monocarboxy isooctyl ester (MCOH) This analyte has not been tested by NHANES
Cyclohexane-1,2-dicarboxylic acid-mono(hydroxy-isononyl) ester (MHNCH) 1.57 3.81 1.39
Last Updated: May 1, 2017