National Health and Nutrition Examination Survey

The primary purpose of the NHANES 2005-2006 allergy component was to investigate the effects of common indoor allergens on allergic sensitization and disease in the non-institutionalized U.S. population. The allergy component consisted of 3 parts. First, a nationally representative sample of household dust was collected to measure the amounts of 10 indoor allergens and endotoxin that were present. Second, a blood sample was also drawn at the NHANES examination site for allergen specific immunoglobin E (IgE) antibody testing to these same allergens (AL_IGE_D). Third, data on self-reported allergic symptoms and conditions and household characteristics was collected during household interview questionnaires (AGQ_D, RDQ_D, MCQ_D, HOQ_D).

The allergens in dust dataset (ALDUST_D), contains results for 10 common indoor allergens and endotoxin from the collected dust sample extract. The 10 specific allergens tested were:

Bla g 1 and Bla g 2 are allergens from German cockroach, typically found in a household’s kitchen and/or bathrooms. Bla g 1 shows antigenic cross-reactivity with American cockroach (Periplaneta americana). Can f 1 is an allergen from dogs. This allergen occurs on particulates that are readily transportable, and can be found throughout the home environment. Fel d 1 is an allergen from cats. Similar to dog antigen, Fel d 1 occurs on particulates that are readily transportable, and can also be found throughout the home environment. Der p 1 and Der f 1 are allergens from house dust mites. In homes, the highest concentrations of dust mites are usually found in mattresses, pillows, and upholstered furniture. Mus m 1 is an allergen from mouse urine. Buildings with exposed food supplies, poor housing maintenance, and difficulties with litter and sanitation provide a conducive environment for mice. Rat n 1 is an allergen from rats which have a similar habitat to mice. Rat allergen is a less commonly identified, yet is an important cause of IgE-mediated hypersensitivity.

The two fungal antigens (Alternaria alternata and Aspergillus fumigatus) measured are primarily outdoor antigens, which are transported into the indoor environment. Alternaria alternata is a species of cosmopolitan dematiaceous fungi commonly isolated from plants, soil, and food. Allergy to Alternaria alternata is one of the more prevalent allergies among mold sensitive atopic persons. Aspergillus fumigatus is a common outdoor mold associated with decaying vegetative matter. It is one of the molds commonly associated with allergic and/or asthmatic symptoms.

Dust endotoxin was also measured with the dust allergens. Endotoxin is an inflammatory lipopolysaccharide from gram-negative bacteria that is found in the indoor environment. Indoor sources of endotoxin include pets, pests, humidifiers, and outdoor air. The inhalation of dust endotoxin has been linked to asthma, allergic rhinitis, and wheezing.

The ALDUST_D dataset was developed to provide data relevant to examining the relationship between specific indoor allergen concentrations and the risk of allergic sensitization, symptoms, and disease. From the participant’s perspective, demonstrating a quantitative relationship between allergen concentrations and allergic outcomes could possibly be used to determine if action is needed to reduce allergen exposures, and hence, subsequently reduce risks to health. The methodology developed in this study could potentially be useful to monitor and verify allergen reduction efforts.

A full sample of NHANES 2005-2006 participants aged 1 year and older, who had been examined in the NHANES Mobile Examination Center (MEC), were eligible for dust allergen testing. There were no exclusions or any precluding conditions for the household dust collection component.

Sample Collection Methodology:

The household dust sample collection procedure is described in detail in the NHANES Allergen Dust Collection Procedures Manual. All allergy household dust collections were performed by trained technicians using a standardized protocol; and informed consent was obtained from all participants sampled. The dust samples collected were a combined bed and bedroom floor sample for each subject. These sampling sites were chosen because it is generally believed that the bedroom is the relevant site for exposure to indoor allergens. A Sanitaire™ Model 3683 vacuum cleaner and a Mitest™ Dust Collector (Indoor Biotechnologies, Inc., Charlottesville, VA) were used for sample collection.

For each sample collected, a 1-square yard surface was marked on both the bed surface and also on the adjacent bedroom floor.  Each site was individually vacuumed for two minutes (a total of 4 minutes vacuum time per sample collection). Specially designed disposable 1-square yard templates (rectangular with the dimensions 2.25 ft. x 4 ft.) were used to mark off areas to be vacuumed. In bed sampling, technicians were instructed to obtain vacuum samples from the surface of the bed sheets wherever possible. In floor sampling, for rooms where there was a choice between carpeted and uncarpeted floor areas, technicians were instructed to place the sample template on the carpet as opposed to a smooth floor surface.

For the purpose of the dust sample collection, a participant’s bedroom was defined as the room or area where that participant regularly slept, and where his/her primary bed (defined below) was located. In the majority of cases, the participant’s bedroom was a separate room; however, in some instances a bedroom area in the home was in a living room, a den, a hallway, an attic, a porch, etc. The participant’s bedroom may also be a bedroom shared with other persons. A “bed” was defined as the place that the participant usually slept. While a bed was most typically a standard bedframe with a mattress, in some cases it was a cot, a crib, a mattress on the floor, a blanket or sleeping bag on the floor, a mattress on a sleeper sofa, a sofa, or a roll away bed. The variable AADBDTYP provides data on the type of bed sampled. In cases where the participant slept in multiple places in the home, the bed where they slept most of the time was chosen for the dust sample collection. In instances where two sample participants slept in the same bed, only a single bed/floor dust sample was obtained and the sample results were assigned equally to both participants.

The standard template could not be used if the available bed surface or floor surface was less than 1-square yard. In such cases, a smaller surface area was taped off with masking tape and vacuumed for a reduced period of time, adjusted to keep the rate of vacuuming equivalent to that of the standard template vacuumed for 2 minutes. Derived variables were created for the vacuumed area and time as follows:

When collected, the dust sample, contained within the Mitest™ Dust Collector, was transported immediately, in a cooler, to the local NHANES field office where it was stored at 4º C until it was shipped to the laboratory.

These additional variables describe the sampling procedure:

Dust Allergen and Endotoxin Laboratory Analysis:

Description of dust sample preparation:

The collected dust specimens were sent to the University of Iowa dust endotoxin laboratory where they were stored at -80C until analyses. The sieved dust samples were weighed ( usually 50 mg) and extracted with 1 mL of sterile pyrogen-free LAL water plus 0.05% Tween 20™ to a final concentration of 50 mg/mL. Prior to analysis, the dust was shaken for 1 hour at room temperature and centrifuged at 4ºC to remove large insoluble particles. The supernatant was used for dust endotoxin analyses. The supernatant was then mixed with pyrogen-free phosphate buffered saline and the mixture was shipped on dry ice to the following dust allergen laboratories: Indoor Biotechnologies, Inc. and Air Quality Sciences, Inc. Detailed specimen collection and processing instructions are discussed in the NHANES Laboratory/Medical Technologists Procedures Manual. Read the LABDOC file for detailed data processing and editing protocols.

Descriptions of Laboratory Methods:

Description of MARIA assay for 8 Allergens:
Household dust samples were analyzed for a panel of 8 allergens using the Indoor Biotechnologies ELISA MARIA ® Multiplex Array assay. Cockroach (Bla g 2), Dog (Can f 1), Cat (Fel d 1), Dust Mites (Der p 1; Der f 1), Mouse (Mus m 1), Rat (Rat n 1) and the fungus Alternaria alternata (Alt a 1) were analyzed by this method. The Multiplex Array for Indoor Allergens (MARIA) is an allergen detection assay developed by Indoor Biotechnologies, Inc. The assay is based on Luminex xMap® technology that utilizes beads with unique ratios of internal fluorescent dyes coupled to allergen-specific antibodies. The beads are mixed with a sample and bind the allergen of interest. A biotinylated detection antibody is added, followed by a streptavidin-conjugated fluorophore. The beads are read by an instrument equipped with lasers to identify each bead set and quantify the fluorescent intensity resulting from the bound allergen. The fluorescence intensity for each allergen is compared to a standard curve to determine the allergen concentration in the sample. The detailed laboratory method for MARIA is in the NHANES Laboratory/Medical Technologists Procedures Manual.

Cockroach Bla g 1 assay:
The Cockroach Bla g 1 assay was performed at Air Quality Sciences, Inc. The Bla g 1 assay used an enzyme-linked immunosorbent assay (ELISA) test kit from Indoor Biotechnologies Inc. The allergen is measured using monoclonal capture and polyclonal detection antibodies. The reaction uses an enzyme conjugated to the detection antibody and generates a colored product. The intensity of the color is proportional to the concentration of the allergen. The concentration is determined by measuring color intensity at a specific wavelength. The detailed laboratory method for Bla g 1 assay is in the NHANES Laboratory/Medical Technologists Procedures Manual.

Aspergillus fumigatus assay:
The Aspergillus fumigatus assay was performed at Air Quality Sciences, Inc. The method measures Aspergillus fumigatus antigens in samples, using a custom-prepared, antiserum enzyme-linked immunosorbent assay (ELISA) test from Greer Laboratories (Lenoir, NC). The reaction uses an enzyme conjugated to the detection antibody and generates a colored product. The intensity of the color is proportional to the concentration of antigen. The concentration is determined by measuring color intensity at a specific wavelength. The detailed laboratory method for Aspergillus fumigatus assay is in the NHANES Laboratory/Medical Technologists Procedures Manual.

Dust endotoxin assay:
The dust endotoxin assay was performed at the University of Iowa laboratory. The dust endotoxin is measured using a Limulus amebocyte lysate (LAL) assay. The LAL assay is based on the sensitivity of an enzymatic clotting cascade in the amebocytes found in the hemolymph of the horse-shoe crab Limulus polyphemus. The dust extracts were measured using a kinetic chromogenic assay. Endotoxin from E. coli 0155 was used in standard curves. Aliquots of endotoxin standards and extract were pipetted into a pyrogen-free microplate and LAL reagent and substrate was added. The absorbance was read at 405 nm. The detailed laboratory method for the endotoxin assay is in the NHANES Laboratory/Medical Technologists Procedures Manual.

During the data editing process, data distributions and outlier values were examined. Since there was not sufficient information to conclude that any specific measurements were invalid, no data were excluded from the data set.

Results and Limit of Detection (LOD):

The DSX______ variable (with the underscore filled by an allergen-specific code) provides the analytic result for that analyte. The detection limits were different for all of the analytes. The variable “DSD___LC” (with the underscore filled by an allergen-specific code) indicates whether the result was either measurable or below the LOD. There are two values: “0” and “1.” A “0” value means that the result was above or at the LOD. A “1” indicates that the result was below the LOD.

The variable DSX___LD provides the value for the limit of detection. Some analytes had multiple limits of detection. For samples below the LOD, fill values equal to the LOD divided by the square root of 2 are used in the DSX___ result field.

The NHANES quality control and quality assurance protocols (QA/QC) meet the 1988 Clinical Laboratory Improvement Act mandates. Detailed quality control and quality assurance instructions are discussed in the NHANES Laboratory/Medical Technologists Procedures Manual.

The ALDUST_D data file includes the laboratory values, dust weights, and other variables that describe characteristics of the bed, floor, and room where the dust samples were collected. Additional administrative variables were created for the dataset as follows:

Status of Dust Allergen Data (AADEXSTS): A final allergen dust component status code was created to provide analysts with an overview of the completeness of the laboratory data for 10 dust allergens and endotoxin.

Whether the participant had moved since the time of the household interview (AADMOVE). This variable flags whether a participant had moved since his/her household interview. When this was the case, data variables related to the participant’s home in the household interview may or may not relate to the participant’s current living situation at the time the allergen dust samples were collected.

Dust allergen and endotoxin amounts per sieved dust weight calculations:
The dust allergens and endotoxin concentrations are provided in the file. The following formulas should be used to calculate dust allergen and endotoxin amounts per sieved dust weight:

Aspergillus fumigatus: (ng antigen/mL)*(1mL/45mg dust)*(1ug antigen/1000 ng antigen)*(1000 mg dust/1 g dust) = ug Aspergillus fumigatus/g dust [multiply by factor of 0.0222 or 1/45]

Cockroach Bla g 1: (U allergen/mL)*(1mL/45 mg dust)*(1000 mg dust/1 g dust) = U Bla g 1/g dust [multiply by factor of 22.22]

MARIA allergen: (ng allergen/mL)*(1mL/45mg dust) *(1000 mg dust/1 g dust) = ng allergen/g dust [multiply by factor of 22.22]

Endotoxin: (EU/mL)*(1ml/50 mg dust) = EU/mg dust [multiply by factor of 0.020 or 1/50]

The fill values for dust allergen and endotoxin concentrations below the LOD should be used for the concentration in the calculations above.

Samples with Valid Dust Allergen and Endotoxin Concentrations but with Missing Sieved and/or Total Dust Weights:

Due to clerical or administrative errors in the data collection file, a small number of participants had missing or illogical sieved dust weight and/or total dust weight values. As these participants had valid reported lab values for the dust allergens and endotoxin, they did in fact have adequate dust sample collection for analysis, with only the specific quantity of collected or sieved dust being unknown. Therefore, the data for dust allergen and endotoxin concentrations were retained in the current data release. 

2-Year MEC examination weights of dust allergen subsample (WTAL2YR):

Separate dust allergen examination subsample weights were created by re-weighting the standard MEC examination weights (WTMEC2YR) to account for selective non-response and any missing laboratory data. For the NHANES 2005-06 dust allergen component (ALDUST_D) only MEC-examined participants aged 1 year and older were eligible.

NHANES data is weighted at the interview and examination level to account for unit non-response and to provide estimates for the U.S. civilian non-institutionalized population. The allergy dust sample took place after the examination and had missing data at two stages. First, some individuals refused or were otherwise unavailable for collection of the dust sample. Additionally, because of difficulties during the laboratory analysis phase of this component a decision was made to retest specimens. When retesting occurred some samples did not have adequate remaining dust sample available to repeat the laboratory analysis. Because of these two sources of non-response, special weights for the allergy dust samples are included in this data set to adjust for non-response and to permit estimates for the U.S. civilian non-institutionalized population to be made. Because weighting class adjustments frequently result in increases in the variance of survey estimates when many weighting classes are created with relatively few respondents in each class, NHANES avoids the creation of adjustment cells with fewer than 30 sample cases. This requirement limited the adjustment cells to those defined by age, race/Hispanic origin and gender. Even with these few adjustment variables two cells fell below the desirable level of 30 cases.

Proper analysis of NHANES data requires the use of weights and the allergy dust subsample weights are provided for the convenience of the user. However, the user may opt to create their own weights to analyze this data. Some considerations follow.

Properly weighted estimates based on the data obtained from a survey will be approximately unbiased if there was no missing data at any stage. This did not occur in NHANES. For certain important characteristics, the respondents may differ significantly from the nonrespondents. If such differences exist and are not adjusted for in the analyses, then any estimates or inferences made to the target population may be misleading.

Nonresponse adjustment methods such as weighting class adjustment can serve to reduce nonresponse bias. However, the total elimination of such bias is not possible, since within any weighting class the respondents ordinarily will not be fully representative of the nonrespondents.

Taking nonresponse into account in data analyses is the analyst’s responsibility. The first step of any nonresponse bias assessment is to identify the covariates that are highly correlated with the variable of interest and the response status. For example, in this specific instance did the volume of dust collected relate to the values of the dust allergens? The analyst should compare the distribution of respondents and nonrespondents for characteristics that are known for both groups. The analyst should determine whether the level of nonresponse for an item is significant, relative to the bias and variance (relative standard error) needed for a particular inference. Also, do the conclusions agree with results from other studies or contradict them? Additionally, the analyst has the responsibility of pointing out the caveats from nonresponse in the reports such as identifying potential effects on the estimates when nonresponse may be high for a specific comparison.

Other variables of interest:

The NHANES 2005-2006 Allergy Dust data collection included an NHANES household interview questionnaire section (AGQ_D) and a laboratory analysis of serum IGE antibodies (AL_IGE_D). The latter included antibody testing for the same set of allergy antigens measured in the household dust samples. The AGQ_D household interview questionnaire provides personal data on common allergic diseases including hay fever, allergies, sinusitis and eczema. It also provides self-reported information on symptoms associated with those conditions and pet avoidance due to allergies or asthma. Questionnaire data on the respiratory symptom history is contained in the RDQ_D dataset, and the history of asthma diagnosis is contained in the MCQ_D dataset. Additionally, data on the participant’s residence is provided in the NHANES 2005-2006 Housing Characteristics dataset (HOQ_D). This provides data on the type of home, the number of apartments in the building, the age of home, the number of rooms in home, time lived in home, whether the home is owned or rented, water source and treatment, and allergy component related questions about the presence of mold, cockroaches, and animals. The Fasting Questionnaire File includes auxiliary information, such as the conditions precluding venipuncture.