An algorithm for quantitatively estimating occupational endotoxin exposure in the Biomarkers of Exposure and Effect in Agriculture (BEEA) study: I. Development of task‐specific exposure levels from published data

Abstract

Background/Objective: Farmers conduct numerous tasks with potential for endotoxin exposure. As a first step to characterize endotoxin exposure for farmers in the Biomarkers of Exposure and Effect in Agriculture (BEEA) Study, we used published data to estimate task-specific endotoxin concentrations. Methods:We extracted published data on task-specific, personal, inhalable endotoxin concentrations for agricultural tasks queried in the study questionnaire.The data, usually abstracted as summary measures, were evaluated using metaregression models that weighted each geometric mean (GM, natural-log transformed) by the inverse of its within-study variance to obtain task-specific predicted GMs.Results: We extracted 90 endotoxin summary statistics from 26 studies for 9 animalrelated tasks, 30 summary statistics from 6 studies for 3 crop-related tasks, and 10 summary statistics from 5 studies for 4 stored grain-related tasks. Work in poultry and swine confinement facilities, grinding feed, veterinarian services, and cleaning grain bins had predicted GMs > 1000 EU/m 3 . In contrast, harvesting or hauling grain and other crop-related tasks had predicted GMs below 100 EU/m 3 .Significance: These task-specific endotoxin GMs demonstrated exposure variability across common agricultural tasks. These estimates will be used in conjunction with questionnaire responses on task duration to quantitatively estimate endotoxin exposure for study participants, described in a companion paper.

An algorithm for quantitatively estimating occupational endotoxin exposure in the biomarkers of exposure and effect in agriculture study: II. Application to the study population

“…The highest cumulative endotoxin exposures, which combined exposure intensity with task duration, were observed for swine confinement , poultry confinement , and grind feed . This was unsurprising as they were also three of the most highly exposed activities identified in our meta‐regression of published data 18 . The first two tasks have also been used as surrogate measures in epidemiologic analyses, such as number of hogs 9 or livestock or poultry 7 .…”

“…The durations related to cleaning and feeding tasks were included in the duration for the broader questions. Furthermore, these specific tasks had similar intensity estimates as the broader tasks 18 . We also omitted tasks related to fruit, ground vegetables, nuts, tobacco, grass, and other miscellaneous crops because the questionnaire provided no information on the farmers' time spent on activities related to these crops.…”

“…We also omitted tasks related to fruit, ground vegetables, nuts, tobacco, grass, and other miscellaneous crops because the questionnaire provided no information on the farmers' time spent on activities related to these crops. However, these were generally lower exposed activities, with predicted endotoxin geometric means (GM) < 100 EU/m 3 for crop‐related tasks 18 . We also omitted tasks related to cotton because we had no exposure measurements characterizing endotoxin exposure for cotton‐related tasks; however, only 6 BEEA participants who completed the 2014 version of the questionnaire reported growing cotton.…”

“…For the 13 selected tasks ( t ), we estimated task‐specific intensity, I t , in units of EU/m 3 from the predicted GM of a meta‐analysis of published personal, inhalable, task‐specific endotoxin measurements 18 (with GMs noted in Table 1). We also extracted task—( t ), time window—( w ), and subject‐specific ( i) durations, D twi from the questionnaire.…”

“…The algorithm was developed to estimate endotoxin exposure for 13 tasks (five in common with the aforementioned glucan assessment 15 ) during four time windows, reflecting agricultural activities conducted in the past day, week, month, and year. A crucial first step was to derive task‐specific estimates based on published endotoxin data for each of the 13 tasks, which was described in detail in a companion paper 18 . Here, we describe how we combined these task‐specific endotoxin estimates with the subject‐specific task durations reported in questionnaires to obtain task‐ and time window‐specific cumulative endotoxin exposure levels for the BEEA Study participants.…”