key: cord-0993385-vjd7b4f0 authors: Coffman, V. R.; Hall, D. J.; Pisanic, N.; Nadimpalli, M.; McCormack, M.; Diener-West, M.; Davis, M. F.; Heaney, C. D. title: The use of personal protective equipment during common industrial hog operation work activities and acute lung function changes in a prospective worker cohort, North Carolina, USA date: 2020-11-04 journal: medRxiv : the preprint server for health sciences DOI: 10.1101/2020.11.03.20205252 sha: 99aefbb3540a93e1332b08a79a596ae4aa2a7ba4 doc_id: 993385 cord_uid: vjd7b4f0 Objectives: As occupational activities related to acute industrial hog operation (IHO) worker lung function are not well defined, we aimed to identify IHO work activities associated with diminished respiratory function and the effectiveness, if any, of personal protective equipment (PPE) on IHOs. Methods: From 2014-2015, 103 IHO workers were enrolled and followed for 16 weeks. At each bi-weekly visit, lung function measurements were collected via spirometry and work activities and PPE use were self-reported via questionnaire. Generalized linear and linear fixed-effects models were fitted to cross-sectional and longitudinal data. Results: At baseline, years worked on an IHO were associated with diminished lung function, but other activities were less consistent in direction and magnitude. In longitudinal models, only reports of working in feeding/finisher barns, showed a consistent association. However, a -0.3 L (95% confidence interval: -0.6, -0.04) difference in FEV1 was estimated when workers wore PPE consistently versus those weeks they did not. In post-hoc analyses, we found that coveralls and facemasks were worn less consistently when workers experienced worse barn conditions and had more contact with pigs, but coveralls were worn more consistently as cleaning activities increased. Conclusions: Similar to past studies, baseline estimates were likely obscured by healthy worker bias. Also making it challenging to disentangle the effect of work activities on lung function was the discovery that IHO workers used PPE differently according to work task. These data suggest that interventions may be targeted toward improving barn conditions so that workers can consistently utilize IHO-provided PPE. for additional research to identify workplace factors that account for the high prevalence of 144 respiratory symptoms in IHO worker populations. [17, 18] studies for fear of termination, limited access to operations, and corporate influence [19] mean studies are often short-term, lack the ability to determine changes to health on a granular 149 level, and are conducted outside the U.S. This lack of updated data hinders priority setting for 150 the occupational health of the ~33,000 IHO workers in the States,[19] many of whom are 151 from marginalized communities (e.g., low socioeconomic status, lacking health insurance, 152 minorities, and immigrants). 153 The work presented here is unique in that by employing fixed-effects regression 154 analyses, [20] workers are compared to themselves. This largely removes confounding from 155 fixed characteristics (e.g., participant age, sex, race/ethnicity, structural differences in barns, 156 perceived dustiness, and at-home characteristics), which may be present in models estimating 157 the difference in lung function between IHO workers and the general public or other 158 agricultural workers. [15, 21] To the best of our knowledge, this is the first time fixed-effects 159 regression has been used to assess the association between spirometry and IHO work 160 activities and PPE use. In addition, prior studies have examined cross-sectional, [ The purpose of this investigation is to examine the effect a variety of modern IHO 164 work activities and personal protective equipment (PPE) have on lung function changes (i.e., 165 spirometric measurements) on the acute scale. 166 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. participants "typically," "usually," or "ever" did on-operation and what environmental 204 conditions they were "typically," "usually," or "ever" exposed to at work. 205 At each follow-up visit, information about the frequency, magnitude, and duration of 206 participants' contact with pigs, job activities, personal behaviors (e.g., cigarette use), and PPE 207 use was collected. Each question was asked about the week prior. For example, "In the past 208 week have you…" 209 210 Assessment of lung function 211 Lung function was assessed by spirometry. A portable Piko-1 pulmonary function 212 device, which records forced expiratory volume in the first second (FEV 1 ) and peak 213 and Health (NIOSH)-spirometry to be on-site. Additionally, a Koko machine was employed including: FEV 1 , PEFr, forced vital capacity (FVC), the ratio of forced expiratory volume in 219 the first second to forced vital capacity (FEV 1 /FVC), forced expiratory volume in six seconds 220 (FEV 6 ), forced expiratory flow at 25-75% of the pulmonary volume (FEF25-75%), and the 221 ratio of forced expiratory volume in the first second to forced expiratory volume in six 222 seconds (FEV 1 /FEV 6 ). Ultimately, in analyses only the Piko-1 measurements and FEV 1 , 223 PEFr, FVC, FEV 1 /FVC Koko measurements were used. Three good trials, as defined by the 224 ATS[ATS] and the NIOSH standards, [30] were attempted for each testing session unless the 225 participant was physically unable to complete the three spirometric maneuvers. The best 226 maneuver per study visit for each person was used in data analysis. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; Spirometry data were analyzed in two ways: (1) all data from the best try (i.e., no 260 cough, inhale, or delay in start, and a good effort) was used; and (2) as a sensitivity analysis 261 for each model, only measurements with three good tries and two repeatable 262 measurements [29, 30] were used (see Supplemental Material). The decision to use all 263 spirometry measurements (regardless of ATS/NIOSH acceptability) as main analyses was 264 based on: (1) Piko-1 devices were not designed to conform to ATS criteria; (2) only the first 265 three maneuvers were recorded using the Piko-1 device (more maneuvers would have met 266 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; https://doi.org/10.1101/2020.11.03.20205252 doi: medRxiv preprint 1 and 90 workers remained enrolled in the final week and provided 87 measurements ( Figure 280 1). In examining key characteristics stratified by those who were able to complete spirometry 281 tests at baseline and those who were not, we saw little difference except that more non-282 smokers than smokers were able to provide Koko maneuvers and overall more participants 283 were able to provide Piko-1 measurements (98 of 103) than Koko measurements (69 of 103) 284 ( Relationships between the most reported symptoms at enrollment and percent 295 predicted spirometry values were examined to assess the need to include symptoms as 296 confounders in subsequent models (Table S2 and S3). Only report of doctor-diagnosed 297 asthma (a static characteristic) was significantly associated with lung function decline and so 298 symptoms were not included. 299 Ten binary self-reported measures of exposure and two variables transformed into 300 tertiles were examined in association with baseline measures of lung function (Table 3) . 301 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; confidence interval did not include the null), with one relationship (FEV 1 /FVC and pesticide 303 application) slightly significant but in the unhypothesized direction. Worse lung function was 304 seen in those who worked on any IHO longer than those in the first tertile, with significant 305 associations for trends in 3 of 8 models (Table 3) . In sensitivity analyses, adjusting for the 306 hour of spirometric test, current cigarette smoking, and interviewer, the relationship between 307 pesticide application and FEV 1 /FVC strengthened (Table S4) , while the associations between 308 length of time working on any IHO weakened, although 14 of 16 remained in the negative 309 direction. 310 In longitudinal analyses, limited changes in respiratory function were observed 311 between weeks when an activity was versus was not performed ( Table 4) . 312 However, PPE use was statistically significant in the un-hypothesized adverse 313 direction in both crude and adjusted models and when exposure was either collapsed into 314 binary measures (Table 4 ) or expanded to detail the number of activities performed (Table 315 S5). 316 To further explore the relationship between PPE use and diminished lung function of 317 this un-hypothesized association, we post-hoc chose to examine the association between type 318 of PPE use and lung function using fixed-effects linear regression (Table S6) Table S6) . 324 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; https://doi.org/10.1101/2020.11.03.20205252 doi: medRxiv preprint Pearson chi-squared tests for associations between on-IHO work activities and PPE use 326 (Table 5) and reported symptoms and PPE use were employed (Table S7) . 327 While symptoms were not meaningfully associated with different levels of PPE 328 utilization, work activities were. Reports of consistent usage of body protection (irrespective 329 of face protection and with non-consistent face protection) as well as the use of both types of 330 PPE increased during weeks when the most cleaning activities were reported ( Table 5) . 331 Notably, body and face protection use (irrespective of the other type of PPE) declined as 332 reports of barn conditions worsened and as contact with pigs increased, as did the use of both 333 types of protection. 334 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; https://doi.org/10.1101/2020.11.03.20205252 doi: medRxiv preprint corroborated in this cohort (Table 2) , also pointing to healthy worker populations on IHOs. 361 362 Longitudinal analyses were largely null 363 Unexpectedly, longitudinal analyses did not produce many meaningful results. This is 364 perhaps because of the time scale chosen (once every two weeks), due to the limited 365 measurements of the devices used, or lung function impairment that is too early to detect, as 366 was seen in a study by Schwartz et al.[40] It is unlikely that there is no true association 367 between IHO exposures and worker lung function as this has been well documented in prior 368 studies. [41] 369 We did observe, however, a -0.3 L estimated difference in FEV 1 during weeks when 370 workers wore any PPE consistently versus those weeks when they did not, which may be 371 explained by (1) improper use leading to a sense of security and thus allowing higher 372 exposures, (2) re-exposure from previously worn, unclean equipment, and/or ( This cohort of 103 IHO workers is a non-random, self-selected group, which may 400 lead to potential selection bias arising from differences in those who participated versus those 401 who did not (e.g., health status, understanding of exposure issues, interest in protecting 402 themselves from work hazards). However, the population is believed to represent the 403 occupational population demographics from the area from which they were enrolled 404 including low-income, minorities, and those living with elder relatives. Further limitations in 405 generalizability include the possible differences between warmer climates (North Carolina) 406 and colder climates, where ventilation systems may differ. Our study was also limited by 407 healthy worker effect bias as detailed above. 408 409 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; https://doi.org/10.1101/2020.11.03.20205252 doi: medRxiv preprint spirometer to capture FEV 1 readings was a strength. This easy-to-use handheld device 412 allowed more measurements from workers than had a gold standard Koko tabletop device 413 been used (Tables 1 and 2) . The Piko-1 was also able to capture more smokers than the 414 Koko. Of the smokers, 12 of 13 had Piko-1 measurements at baseline, while only 5 of 13 had 415 Koko ( Table 1 ). Considering that the Piko-1 is not a NIOSH gold standard instrument, it is 416 worth noting that when ATS/NIOSH acceptability criteria are applied to these measurements, 417 the resulting point estimates and confidence intervals were similar to those without these 418 criteria applied (data not shown). In this study we found that healthy IHO workers in North Carolina, the longer a 428 person worked on any IHO the worse their lung function became, as measured by spirometry. 429 It also showed that quality FEV 1 data can be collected through the use of a handheld, portable 430 asthma-tracking device that does not require a NIOSH-trained technician to be on-site to 431 operate. Further, PPE usage was associated with declines in the respiratory health of IHO 432 workers but is believed to be strongly associated with work activities, where people use face 433 protection less consistently as they perform tasks that involve increased contact with pigs and 434 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; worsening barn conditions, but wear coveralls more consistently when applying chemicals at determine what kinds of masks and coveralls are being used, how they are cleaned, what 437 guidance is being provided to workers, and if interventions in barn conditions could increase 438 compliance and help protect lung function. 439 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; https://doi.org/10.1101/2020.11.03.20205252 doi: medRxiv preprint This study would not have been possible without a strong partnership between 441 researchers and North Carolinian community-based organization members who have fostered 442 the trust of too often marginalized and at-risk community members. The authors deeply thank 443 the workers who participated in this study. 444 This manuscript is dedicated to the memory of Dr. Steve Wing, who helped conceive 445 the design and analytical framework for this cohort study. 446 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; https://doi.org/10.1101/2020.11.03.20205252 doi: medRxiv preprint Due to the exceedingly sensitive nature of this data, it will not be made publicly 448 available. For questions regarding access, please contact Dr. Christopher Heaney 449 cheaney1@jhu.edu. 450 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; https://doi.org/10.1101/2020.11.03.20205252 doi: medRxiv preprint All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; 458 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; Table 3 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted November 4, 2020. ; https://doi.org/10.1101/2020.11.03.20205252 doi: medRxiv preprint Quantification of odors and odorants from swine 517 operations in North Carolina Determinants of dust and endotoxin exposure of 521 pig farmers: development of a control strategy using empirical modelling Respiratory health status in swine producers relates 525 to endotoxin exposure in the presence of low dust levels Endotoxin exposure as a major 529 determinant of lung function decline in pig farmers Occurrence of MRSA in air and housing environment 533 bioaerosol diversity in the standard swine confinement buildings, and assessment of the 538 seasonal effect Airborne multidrug-resistant bacteria isolated from a 542 concentrated swine feeding operation Determinants of longitudinal changes in 546 spirometric function among swine confinement operators and farmers Respiratory dysfunction in swine production 550 facility workers: Dose response relationships of environmental exposures and pulmonary 551 function High prevalence of obstructive lung NAICS 112210 Hog and pig farming, All Counties 2019 Fourth Quarter Fixed effects regression models Environmental and health studies in swine 572 confinement buildings Respiratory symptoms and lung function among 576 workers in swine confinement buildings: a cross-sectional epidemiological study Respiratory symptoms, lung function, and 581 IgG4 levels against pig antigens in a sample of Dutch pig farmers Working in swine confinement buildings causes an accelerated 589 decline in FEV1: A 7 yr follow up of Danish farmers Longitudinal evaluation of dose response 593 relationships for environmental exposures and pulmonary function in swine production 594 workers Accelerated lung function decline in 597 swine confinement workers Standardisation of spirometry Spirometry quality assurance; common errors and 608 their impact on test results Staphylococcus aureus nasal carriage and recent skin and soft tissue infection among 612 industrial hog operation workers Spirometric reference values from a sample of 616 the general US population Diurnal variations in human pulmonary function Global Initiative for Chronic Obstructive Lung Disease 623 (GOLD) guidelines for chronic obstructive pulmonary disease Bender 627 JB. Personal protective equipment use and handwashing among animal farmers: A multi-site 628 assessment Assessment of personal protective 632 equipment use among Midwestern farmers Awareness and practices 636 regarding zoonotic influenza prevention in Romanian swine workers. Influenza Other Respir 637 Concentrated swine feeding operations and public health: a 640 review of occupational and community health effects Lung function and farm size predict 644 healthy worker effect in swine farmers Airway injury in swine confinement workers Lung function studies in farmers during work in swine confinement 652 units Positive human health effects of wearing 656 responses following exposure in a swine house Respiratory manifestations and respiratory diseases: 664 prevalence and risk factors among pig farmers in Braço do Norte, Brazil Characterization of influenza A 668 outbreaks in Minnesota swine herds and measures taken to reduce the risk of zoonotic 669 transmission Interaction of the role of concentrated animal feeding operations (CAFOs) 673 in emerging infectious diseases (EIDS). Infection COVID-19 exposes animal agriculture's vulnerability. Agriculture and Human 677 Values No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted cohort, North Carolina, 2013-2014 using GLM and clustered at the household level. 460