Search

Despite the prominent use of complex survey data and the growing popularity of machine learning methods in epidemiologic research, few machine learning software implementations offer options for handling complex samples. A major challenge impeding the broader incorporation of machine learning into epidemiologic research is incomplete guidance for analyzing complex survey data, including the importance of sampling weights for valid prediction in target populations. Using data from 15, 820 participants in the 1988-1994 National Health and Nutrition Examination Survey cohort, we determined whether ignoring weights in gradient boosting models of all-cause mortality affected prediction, as measured by the F1 score and corresponding 95% confidence intervals. In simulations, we additionally assessed the impact of sample size, weight variability, predictor strength, and model dimensionality. In the National Health and Nutrition Examination Survey data, unweighted model performance was inflated compared to the weighted model (F1 score 81.9% [95% confidence interval: 81.2%, 82.7%] vs 77.4% [95% confidence interval: 76.1%, 78.6%]). However, the error was mitigated if the F1 score was subsequently recalculated with observed outcomes from the weighted dataset (F1: 77.0%; 95% confidence interval: 75.7%, 78.4%). In simulations, this finding held in the largest sample size (N = 10,000) under all analytic conditions assessed. For sample sizes

Consistent performance of allergen assays is essential to ensure reproducibility of exposure assessments for investigations of asthma and occupational allergic disease. This study evaluated intra- and inter-laboratory reproducibility of a fluorescent multiplex array, which simultaneously measures eight indoor allergens in a single reaction well. A multi-center study was performed in nine laboratories in the US and Europe to determine the inter-laboratory variability of an 8-plex array for dust mite, cat, dog, rat, mouse and cockroach allergens. Aliquots of 151 dust extract samples were sent to participating centers and analyzed by each laboratory on three separate occasions. Agreement within and between laboratories was calculated by the concordance correlation coefficient (CCC). Results were obtained for over 32,000 individual allergen measurements. Levels covered a wide range for all allergens from below the lower limit of detection (LLOD = 0.1-9.8 ng/ml) to higher than 6800 ng/ml for all allergens except Mus m 1, which was up to 1700 ng/ml. Results were reproducible within as well as between laboratories. Within laboratories, 94% of CCC were ≥ 0.90, and 80% of intra-laboratory results fell within a 10% coefficient of variance (CV%). Results between laboratories also showed highly significant positive correlations for all allergens (~0.95, p

Exposure to endotoxin in home environments has become a key issue in asthma and allergy research. Most studies have analyzed floor or mattress dust endotoxin, but its validity as a proxy for airborne exposure is unknown, while active airborne dust sampling is not feasible in large-scale population studies because of logistic and financial limitations. We therefore developed and evaluated a simple passive airborne dust collection method for airborne endotoxin exposure assessment. We explored an electrostatic dust fall collector (EDC), consisting of a 42- by 29.6-cm-sized folder with four electrostatic cloths exposed to the air. The EDC was tested during two 14-day periods in seven nonfarm and nine farm homes and in farm stables. In parallel, active airborne dust sampling was performed with Harvard impactors and floor dust collected by vacuuming, using nylon sampling socks. The endotoxin levels could be measured in all EDC cloth extracts. The levels (in EU/m 2 ) between EDCs used simultaneously or in different sampling periods in the same home correlated strongly ( r > 0.8). EDC endotoxin also correlated moderately to strongly ( r = 0.6 to 0.8) with the endotoxin measured by active airborne dust sampling and living room floor dust sampling and—in farm homes—with the endotoxin captured by the EDC in stables. In contrast, endotoxin levels measured by floor dust sampling showed only a poor correlation with the levels measured by active airborne dust sampling. We therefore conclude that measuring endotoxin levels with the EDC is a valid measure of average airborne endotoxin exposure, while reproducibility over time is at least equivalent to that of reservoir dust analyses.

Endotoxin exposure occurs in homes and occupational environments and is known to cause adverse health effects. In order to compare results from different studies and establish standards, airborne endotoxin exposures should be assessed using standardized methods. Although the European Committee for Standardization (CEN) developed guidelines for endotoxin exposure assessment, these leave room for individual interpretation. The influence of methods of sampling, extraction, and analysis has never been investigated in a full experimental design. Thus, we sought to fully elucidate the importance of all facets of endotoxin assessment. Inhalable dust samples collected simultaneously were used to investigate the effects on and interactions with airborne endotoxin concentration in two working environments of filter type (glass fiber or Teflon), transport conditions (with/without desiccant), sample storage (-20 or 4°C), extraction solution (pyrogen-free water [PFW] or PFW plus 0.05% Tween 20), extract storage (-20 or 4°C), and assay solution (PFW or PFW plus 0.05% Tween 20). Four hundred samples were collected and randomly distributed over the 20 combinations of treatments. There were no differences found for transport conditions and storage temperature of extracts. Also, no interactions between study variables existed. Sampling on glass-fiber filters, storage of samples in the freezer, and extraction in PFW plus 0.05% Tween 20 resulted in 1.3-, 1.1-, and 2.1-fold-higher estimated endotoxin concentrations, respectively. Use of PFW plus 0.05% Tween 20 in the assay solution had an additive effect. Thus, this study investigated gaps in the CEN protocol and provides data with which to fully specify a protocol for standardization of endotoxin exposure assessment. Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Inflammatory airway responses to bioaerosols and to their active compounds, such as endotoxin and β(1 → 3)- glucan, vary between individuals. These differences may be explained by variation in cytokine responsiveness, which can be assessed by in vitro stimulation tests with isolated blood leukocytes or lung macrophages. In large- scale population studies, ex vivo induced cytokine production may also be tested with a more simple `whole blood assay’ (WBA). However, applicability of a WBA to characterize a subject’s responsiveness depends largely on its reproducibility. This study was conducted to: 1) assess the within- and between-subject variability in cytokine production in a WBA after stimulation with endotoxin or β(1 → 3)-glucan; and 2) to determine under which conditions this test is most discriminating between subjects and most reproducible within subjects. Blood was collected from 14 healthy volunteers, of whom 10 also participated on a second occasion. Each blood sample was used in two WBA tests; the first WBA was initiated two hours and the second 26 hours after venapuncture. The WBA test itself comprised overnight incubation with serial dilutions of endotoxin [lipopolysaccharide (LPS)] and curdlan (a β(1 → 3)-glucan), after which blood cell supernatant was collected. Interleukin(IL)-1, IL6, IL8 and tumor necrosis factor (TNF) were determined in the supernatant. In all individuals, a dose-dependent production of cytokines was observed for both LPS and curdlan. For all cytokines, variation between subjects was higher than within subjects, and this was most pronounced for IL1 and IL6. There was moderate-to-high correlation in the induced release of all four cytokines, and between cytokine release induced by LPS or curdlan. Optimal stimulation concentrations were 6.25 and 12.5 ng/mL for endotoxin and 12 500 and 25 000 ng/mL for curdlan. Cytokine production in WBA initiated 26 hours after venapuncture showed lower between-subject and larger within-subject variance, thus favoring an early initiation of the assay. In conclusion, measuring endotoxin-or glucan-induced cytokine production in a WBA initiated within two hours after venapuncture appears to be an effective method to determine a person’s cytokine responsiveness, at least in healthy naive subjects. Toxicology and Industrial Health 2002; 18: 15-27.

β-Glucans are components of fungal cell walls and potent stimulants of innate immunity. The majority of research on biological activities of glucans has focused on β-(1→3)-glucans, which have been implicated in relation to fungal exposure-associated respiratory symptoms and as important stimulatory agents in anti-fungal immune responses. Fungi-and bacteria and plants-produce a wide variety of glucans with vast differences in the proportion and arrangement of their β-(1→3)-, -(1→4)- and -(1→6)-glycosidic linkages. Thus far, the pro-inflammatory potential of different β-glucans has not been studied within the same experimental model. Therefore, we compared the potency of 13 different glucan preparations to induce in vitro production of IL-1β, IL-6, IL-8 and TNF-α in human, whole blood cultures. The strongest inducers of all cytokines were pustulan [β-(1→6)-glucan], lichenan [β-(1→3)-(1→4)-glucan], xyloglucan [β-(1→4)-glucan] and pullulan [α-(1→4)-(1→6)-glucan]. Moderate-to-strong cytokine production was observed for curdlan [β-(1→3)-glucan], baker's yeast glucan [β-(1→3)-(1→6)-glucan] and barley glucan [β-(1→3)-(1→4)-glucan], while all other glucan preparations induced very low, or no, detectable levels of cytokines. We therefore conclude that innate immunity reactions are not exclusively induced by β-(1→3)-glucans, but also by β-(1→6)- and β-(1→4)-structures. Thus, not only β-(1→3)-glucan, but also other β-glucans and particularly β-(1→6)-glucans should be considered in future research.

We recently introduced a passive dust sampling method for airborne endotoxin and glucan exposure assessment-the electrostatic dustfall collector (EDC). In this study, we assessed the effects of different storage and extraction procedures on measured endotoxin and glucan levels, using 12 parallel EDC samples from 10 low exposed indoor environments. Additionally, we compared 2- and 4-week sampling with the prospect of reaching higher dust yields. Endotoxin concentrations were highest after extraction with pyrogen-free water (pf water) + Tween. Phosphate-buffered saline (PBS)-Tween yielded significantly (44%) lower levels, and practically no endotoxin was detected after extraction in pf water without Tween. Glucan levels were highest after extraction in PBS-Tween at 120 degrees C, whereas extracts made in NaOH at room temperature or 120 degrees C were completely negative. Direct extraction from the EDC cloth or sequential extraction after a preceding endotoxin extraction yielded comparable glucan levels. Sample storage at different temperatures before extraction did not affect endotoxin and glucan concentrations. Doubling the sampling duration yielded similar endotoxin and only 50% higher glucan levels. In conclusion, of the tested variables, the extraction medium was the predominant factor affecting endotoxin and glucan yields.

Associations between house dust-associated beta-(1,3)-glucan exposure and airway inflammatory reactions have been reported, while such exposures in early childhood have been suggested to protect against asthma and wheezing. Most epidemiological studies have used reservoir dust samples and an inhibition enzyme immunoassay (EIA) for beta-(1,3)-glucan exposure assessment. The objective of this study was to develop inexpensive but highly sensitive enzyme immunoassays to measure airborne beta-(1,3)-glucans in low-exposure environments, like homes. Specificities of available anti-beta-(1,3)-glucan antibodies were defined by direct and inhibition experiments. Three suitable antibody combinations were selected for sandwich EIAs. beta-(1,3)-Glucans in passive airborne dust collected with an electrostatic dust fall collector (EDC) and floor dust from seven homes were measured with the three EIAs. Floor dust samples were additionally analyzed in the inhibition EIA. The sandwich EIAs were sensitive enough for airborne glucan measurement and showed different specificities for commercial glucans, while the beta-(1,3)-glucan levels in house dust samples correlated strongly. The feasibility of measuring glucans in airborne dust with the recently introduced EDC method was further investigated by selecting the most suitable of the three EIAs to measure and compare beta-(1,3)-glucan levels in the EDC and in floor and actively collected airborne dust samples of the previously performed EDC validation study. The EDC beta-(1,3)-glucan levels correlated moderately with beta-(1,3)-glucans in actively collected airborne dust and floor dust samples, while the glucan levels in the airborne dust and floor dust samples did not correlate. The combination of the newly developed beta-(1,3)-glucan sandwich EIA with EDC sampling now allows assessment in large-scale population studies of exposure to airborne beta-(1,3)-glucans in homes or other low-exposure environments.

Toxic gases, vapors, and particles are emitted from concentrated animal feeding operations (CAFOs) into the general environment. These include ammonia, hydrogen sulfide, carbon dioxide, malodorous vapors, and particles contaminated with a wide range of microorganisms. Little is known about the health risks of exposure to these agents for people living in the surrounding areas. Malodor is one of the predominant concerns, and there is evidence that psychophysiologic changes may occur as a result of exposure to malodorous compounds. There is a paucity of data regarding community adverse health effects related to low-level gas and particulate emissions. Most information comes from studies among workers in CAFO installations. Research over the last decades has shown that microbial exposures, especially endotoxin exposure, are related to deleterious respiratory health effects, of which cross-shift lung function decline and accelerated decline over time are the most pronounced effects. Studies in naïve subjects and workers have shown respiratory inflammatory responses related to the microbial load. This working group, which was part of the Conference on Environmental Health Impacts of Concentrated Animal Feeding Operations: Anticipating Hazards-Searching for Solutions, concluded that there is a great need to evaluate health effects from exposures to the toxic gases, vapors, and particles emitted into the general environment by CAFOs. Research should focus not only on nuisance and odors but also on potential health effects from microbial exposures, concentrating on susceptible subgroups, especially asthmatic children and the elderly, since these exposures have been shown to be related to respiratory health effects among workers in CAFOs. Udgivelsesdato: 2007-Feb

Endotoxin exposure is associated with wheeze and asthma morbidity, while early life exposure may reduce risk of allergy and asthma. Unfortunately, it is difficult to compare endotoxin results from different laboratories and environments. We undertook this study to determine if lipopolysaccharide (LPS) extraction efficiency could account for differences among laboratories. We generated and collected aerosols from chicken and swine barns, and corn processing. We randomly allocated side-by-side filter samples to five laboratories for Limulus assay of endotoxin. Lyophilized aliquots of filter extracts were analyzed for 3-hydroxy fatty acids (3-OHFAs) as a marker of LPS using gas chromatography-mass spectrometry. There were significant differences in endotoxin assay and GC-MS (LPS) results between laboratories for all dust types (p < 0.01). Patterns of differences between labs varied by dust type. Relationships between assay and GC/MS results also depended on dust type. The percentages of individual 3-OHFA chain lengths varied across labs (p < 0.0001) suggesting that each lab recovered a different fraction of the LPS available. The presence of large amounts of particle associated LPS and absence of a freezing thawing cycle were associated with lower correlations between LPS and bioactivity, consistent with an absence of Limulus response to cell-bound endotoxin. These data suggest that extraction methods affect endotoxin measurements. The LAL methods may be most suitable when comparing exposures within similar environments; GC-MS offers additional information helpful in optimizing sample treatment and extraction. GC-MS may be of use when comparing across heterogeneous environments and should be considered for inclusion in future studies of human health outcomes.

Exposures to bioaerosols in the occupational environment are associated with a wide range of health effects with major public health impact, including infectious diseases, acute toxic effects, allergies and cancer. Respiratory symptoms and lung function impairment are the most widely studied and probably among the most important bioaerosol-associated health effects. In addition to these adverse health effects some protective effects of microbial exposure on atopy and atopic conditions has also been suggested. New industrial activities have emerged in recent years in which exposures to bioaerosols can be abundant, e.g. the waste recycling and composting industry, biotechnology industries producing highly purified enzymes and the detergent and food industries that make use of these enzymes. Dose–response relationships have not been established for most biological agents and knowledge about threshold values is sparse. Exposure limits are available for some contaminants, e.g. wood dust, subtilisins (bacterial enzymes) and flour dust. Exposure limits for bacterial endotoxin have been proposed. Risk assessment is seriously hampered by the lack of valid quantitative exposure assessment methods. Traditional culture methods to quantify microbial exposures have proven to be of limited use. Non-culture methods and assessment methods for microbial constituents [e.g. allergens, endotoxin, β(1→3)-glucans, fungal extracellular polysaccharides] appear more successful; however, experience with these methods is generally limited. Therefore, more research is needed to establish better exposure assessment tools and validate newly developed methods. Other important areas that require further research include: potential protective effects of microbial exposures on atopy and atopic diseases, inter-individual susceptibility for biological exposures, interactions of bioaerosols with non-biological agents and other potential health effects such as skin and neurological conditions and birth effects.

Bioaerosols are defined as aerosols or particulate matter of microbial, plant or animal origin. Bioaerosols may consist of pathogenic or non-pathogenic live or dead bacteria and fungi, viruses, allergens, bacterial endotoxins, mycotoxins, peptidoglycans, ß(1 -3)-glucans, pollens, plant fibres, erc. Exposures to bioaerosols in the occupational environment are associated with a wide range ofhealth effects, including infectious diseases, toxic effecrs, allergies, and cancer [Douwes er aJ.,2003]. 'W'orkers from a large number of industries are potentially at risk including workers in agriculture, meat production, food and animal feed industry waste recycling and composting industr¡ detergent industr¡ wood and paper industr¡ metal machining industries, biotechnology industries, the medical and public health sector, as well as, veterinarians, pet shop keepers, laboratory animal workers, etc.

In this paper pitfalls in risk assessment for high molecular weight allergens, which can cause typical Type I/IgE-mediated respiratory allergy, are discussed. The major pitfalls seem to be that no agreement exists on the preferential end point that should be used in risk assessment. As a result, it is unclear which exposure–response relationship should be considered. In addition, there is a lack of data on health risks for non-occupationally exposed reference populations, so the baseline risk is often not known and little is known about the shape of exposure–response relationships and the existence of exposure thresholds. The good news is that more and more groups have published exposure–response relationships for several allergens. The possibilities for risk assessment approaches that should lead to occupational exposure standards are explored. Specific consideration is given to situations in which data on exposure–response relationships for humans are available. Considerable progress has been made in this area by use of advanced statistical techniques for exposure–response modelling. The major practical constraint at this moment seems to be the absence of well-standardized measurement techniques (immunoassays) for the evaluation of allergen exposure in the field.

Inflammatory airway responses to bioaerosols and to their active compounds, such as endotoxin and β(1 → 3)- glucan, vary between individuals. These differences may be explained by variation in cytokine responsiveness, which can be assessed by in vitro stimulation tests with isolated blood leukocytes or lung macrophages. In large- scale population studies, ex vivo induced cytokine production may also be tested with a more simple `whole blood assay’ (WBA). However, applicability of a WBA to characterize a subject’s responsiveness depends largely on its reproducibility. This study was conducted to: 1) assess the within- and between-subject variability in cytokine production in a WBA after stimulation with endotoxin or β(1 → 3)-glucan; and 2) to determine under which conditions this test is most discriminating between subjects and most reproducible within subjects. Blood was collected from 14 healthy volunteers, of whom 10 also participated on a second occasion. Each blood sample was used in two WBA tests; the first WBA was initiated two hours and the second 26 hours after venapuncture. The WBA test itself comprised overnight incubation with serial dilutions of endotoxin [lipopolysaccharide (LPS)] and curdlan (a β(1 → 3)-glucan), after which blood cell supernatant was collected. Interleukin(IL)-1, IL6, IL8 and tumor necrosis factor (TNF) were determined in the supernatant. In all individuals, a dose-dependent production of cytokines was observed for both LPS and curdlan. For all cytokines, variation between subjects was higher than within subjects, and this was most pronounced for IL1 and IL6. There was moderate-to-high correlation in the induced release of all four cytokines, and between cytokine release induced by LPS or curdlan. Optimal stimulation concentrations were 6.25 and 12.5 ng/mL for endotoxin and 12 500 and 25 000 ng/mL for curdlan. Cytokine production in WBA initiated 26 hours after venapuncture showed lower between-subject and larger within-subject variance, thus favoring an early initiation of the assay. In conclusion, measuring endotoxin-or glucan-induced cytokine production in a WBA initiated within two hours after venapuncture appears to be an effective method to determine a person’s cytokine responsiveness, at least in healthy naive subjects. Toxicology and Industrial Health 2002; 18: 15-27.