Sampling Locations and Weapons
Sampling was completed at a fully enclosed small arms firing range at WPAFB in Dayton, Ohio and a partially enclosed small arms range at JBC in Charleston, South Carolina. These bases were selected to assess exposure at small arms ranges with different types of ventilation systems. At the fully enclosed range at WPAFB, a perforated ventilation plenum at the top of the back wall provides air flow into, across the ready/firing line, and downrange towards the bullet trap. The plenum delivers air provided by two air handling systems, each providing airflow for one half of the range. All areas are located indoors, and the ventilation system is designed to keep the firing range at a slightly negative pressure. The tower where an instructor provides instruction to the class, is physically separated from the range, and climate control in this room is provided by a separate air handling system. Observation of the firing line is made through bullet-proof windows.
Unlike the small arms range at WPAFB, the partially enclosed range at JBC utilizes a ventilation system comprised of individual, tube-axial ventilation fans. The ventilation system consists of three banks of ten fans each, located just under the roof deck at the back, middle, and end of the range above the bullet trap. This orientation of fans is intended to move air from the firing line toward the bullet trap. The firing line in the partially enclosed range is open to the outdoors, and the roof above the bullet trap is composed of separated pieces, allowing emissions to exhaust out of the range. Also, unlike at WPAFB, the tower is not isolated from the range in a separate room, thus the tower instructor (located just behind the ready line) is also exposed to firing emissions.
At both bases, instructors stand between a yellow ready line, roughly six feet removed from the red firing line where shooters stand to fire their weapons during training classes. For this study, the range instructors attempted to maximize class numbers. Weapons included the M9 pistol, a single-round, semiautomatic weapon with a barrel length of 125 millimeters, and the M4 rifle, which can be operated in single-round or three-round burst mode and has a barrel length of 370 millimeters. M4 and M4/M9 combined training classes fired the rifle in both modes. Both weapons utilized LFF ammunition.
During the study at WPAFB, ventilation assessments were performed daily prior to firing classes. To account for the differences in ranges and ventilation system design, two different approaches were used to measure airflow velocity and direction. Nine-point measurements, in a grid-like fashion, were taken within each stall across the firing line and at the walls using three Shortridge Instruments Air Data Multimeters model 880-C using the Velgrid attachment. These instruments were used as opposed to standard hot wire anemometers because they indicate direction of flow in addition to velocity. The Velgrids average the air velocity across 16 points in a 12 × 12-inch square. The nine-point measurements consisted of three readings each on the left, middle, and right side of each stall at heights of 1 foot (ft), 3 ft, and 5 ft. In order to make the measurements time effective, a cart was constructed which allowed for simultaneous capture of readings at each height.
Using the same measurement technique, ventilation measurements were made at JBC prior to three of the four days of firing due to time constraints. Additionally, wind speed and direction were measured during the M9/M4 combined class and M4 class in order study the effect of wind on emissions. Four Kestrel 5500 Pocket Weather Trackers (Nielsen-Kellerman) were placed two meters behind the instructing line, measuring wind speed and direction with a frequency of 1 Hz. The Kestrels were placed behind lanes 1 (range left), 8 (tower left), 13 (tower right), and 20 (range right). Only three of the four weather trackers recorded data, and as such results are shown with the three locations recorded.
A Degrees Control Inc. C-Breeze artificial fog generator was used to qualitatively measure flow direction at both ranges by walking across the firing line at each stall (1, 9) (Fig. S1 and Fig. S2).
Participant Inclusion and Data Collection
Only active-duty AF members were eligible to participate in the study. Range instructors were recruited from Security Forces at both WPAFB (n = 9) and JBC (n = 4). However, due to reassignments, attendance, and underlying health issues, two instructors from WPAFB and one instructor from JBC were excluded. Additionally, control participants were recruited from Security Forces at each base (WPAFB, n = 5; JBC, n = 5). Across both bases, a total of 11 instructors and 10 controls were included during analysis of survey and biomarker data. All participants were assigned a unique identifier under which their data and samples were logged.
Range instructors at WPAFB were monitored during their workday from 08 Mar 2019 to 20 Mar 2019, and control participants were monitored from 29 Mar 2019 to 10 Apr 2019. The study at JBC took place on 08 Jul 2019 through 19 Jul 2019 with instructors being monitored the first week and controls the second week.
Monitoring included daily questionnaires and a pre-study initial questionnaire, real-time personal breathing zone measurements for metals, CO, and UFPs during the work shift and physiological lung function measurements (pulmonary function test) and biological sample collection (urine) before and after the work shift. This study was reviewed by the Air Force Research Laboratory (ARFL) institutional review board (IRB) and assigned a “not human subjects research” designation (FWR20190025N).
A questionnaire series was constructed on the Survey Monkey platform to capture symptom type and frequency along with lifestyle factors to support interpretation of environmental and biological data. The series consisted of an initial one-time background questionnaire, a daily pre-work questionnaire, and a daily post-work questionnaire. Content design was informed by prior studies of AF Combat Arms populations (1) along with literature on potential biomarker confounders related to lifestyle and health factors (36–40). Occupational health, medical, and range experts reviewed the questionnaires before they were finalized. Additionally, phrasing was adjusted to align with each base’s operations and tempo.
All questionnaires were administered electronically via Apple iPad tablets connected to the Survey Monkey mobile application, and responses were logged under the participant’s unique identifier. On the first day of sampling, participants responded to the background questionnaire upon arriving at the workplace, lasting up to 25 minutes. It captured data related to demographics (age, race, sex, height, weight, rank); lifestyle factors (tobacco use, hobbies, off-duty ammunition use); health status (allergies, medications/supplements, hereditary conditions, dietary factors); occupational history (years in current role, years at base, past duties); past symptom experiences; and current state (exposures within 24 h, active illness or symptoms). Recall timeframes for behavior and health history questions were mostly within the past week while symptomology was within the last month. Due to time constraints, collection of these data topics on the first day at JBC was split between the initial pre- and post-shift questionnaires.
Subsequent pre-shift and post-shift questionnaires were shorter, lasting up to 5 minutes. These questionnaires were administered daily before and after work shifts, often in conjunction with biological sample collection. Content and administration of all questionnaires for control populations mirrored that of instructors, minus the collection of firing range details. Pre-shift questionnaires captured off-duty behaviors since the prior shift, recent consumption of products associated with inflammatory markers (caffeine, tobacco, medication, etc.), symptoms experienced since prior shift, and range assignment (instructors only). Post-shift questionnaires captured duty details from that day and any symptoms experienced during shift. Responses were exported from Survey Monkey to Excel at the end of the sampling period for data management and visualization. Each participant’s responses were linked to environmental and biological sample data by the individual’s unique identifying code for analysis.
Each participant (instructor and control) was issued a personal exposure vest (PEV) prior to beginning their workday and the vest was collected at the end of each workday. PEVs included four pieces of equipment to characterize exposure, including two air sampling pumps connected to filter cassettes to collect particulates for offline analysis of metals, a direct reading gas meter, and a direct reading UFP sampler.
Particulates were collected and analyzed offline for metals using NIOSH 7303. Briefly, two 37 mm MCE filter cassettes were attached to the lapel on each vest and connected to the air sampling pump via Tygon® tubing. One sample was collected for total metals and a second cassette was outfitted with an aluminum cyclone filter (Zefon International) to capture respirable particulate (cut size of 4 µm). Air was pulled through the MCE cassettes via GilAir (Sensidyne) personal sampling pumps with flowrates at 2.5 liters per minute. Pumps were calibrated using a Defender (DryCal®) primary flow calibrator within 5% of desired flowrates before and after each sampling event. If post-calibration of pumps exceeded a 5% difference from pre-calibration, sample volumes were adjusted to reflect the new calibration. The filters were analyzed for metals by ALS Environmental using inductively coupled plasma mass spectrometry (ICP-MS).
CO was monitored by a MultiRAE Pro (RAE Technologies) and data was logged every second for the duration of the exposure. UFPs were monitored using the Partector® (Naneos). This device computes the exposure metric known as the lung deposited surface area (LDSA), which is a measure of the concentration of an aerosol that will deposit in the lung based on the size of the particles (41). The Partector measures LDSA from 0–12,000 µm²/cm³ with an average particle diameter < 300 nm. It also estimates particle number concentration from 0–1x10⁶ particles per cm³ of air.
During firing events, instructor locations were noted to help understand how personal exposure data related to ventilation assessment data. The control population was sampled in the same manner as instructors, though specific location notes were not taken by observers. Location logs were given to the control subjects, so that they could identify their location over the course of the day to explain any anomalies in exposure data.
Pulmonary Function Testing
Individual forced vital capacity (FVC) and first second expiratory volume (FEV1were measured by spirometry using the Spirodoc (Medical International Research) according to manufacturer instructions. Pulmonary measurements were made pre-shift and post-shift. Post-shift measurements were compared to pre-shift measurements and analyzed for significant differences.
Urine Collection and Analysis
Urine was collected pre-shift and post-shift from both instructors and controls every workday. Upon arrival to the collection area, participants were provided with a urine specimen cup prelabeled with their unique identifier. Participants provided a clean catch urine sample which was immediately placed into a secondary Ziploc® bag and stored on ice until transport to the lab. At the lab, a 5 mL aliquot of each bulk urine sample was transferred to a 15 mL conical tube that was prelabeled with the participant unique identifier, date and pre- or post- collection time. Urine samples were centrifuged at 2000 × g for 15 minutes at 4°C and the resulting supernatant was aliquoted into 1 mL volumes for analysis or stored at -80°C.
Levels of urinary 8-OHdG were measured by a commercial enzyme linked immunosorbent assay (ELISA) kit (R&D Systems) according to manufacturer instructions. Briefly, urinary supernatant was diluted 1:10 and 25 microliters (µL) of the diluted sample was loaded per well in duplicate in a 96-well plate. A standard curve was run on each plate. To calculate the concentration of 8-OHdG, the standard curve values were log transformed, plotted, and fit with a polynomial regression line. 8-OHdG levels were normalized to uCr, which was measured using both commercial (Arbor Assays, R&D Systems) and in-house colorimetric assays. The same urine samples used for 8-OHdG analysis were diluted 1:20 in distilled/deionized water and 50 µL of sample was run per well in duplicate. Similar to the 8-OHdH assay, a standard curve was run on each plate. An alkaline picric acid solution was generated by combining 2.5 mL of 1 N NaOH with 12.5 mL of 0.13% Picric Acid and 100 µL was added to each well. After 45 minutes, the absorbance was read at 490 nm. The standard curve was log transformed and fitted with a linear regression line in order to calculate uCr levels. 8-OHdG levels were divided by the uCr levels for normalization. All individual post-shift 8-OHdG/uCr levels were then normalized to the initial pre-shift level (Monday pre) as a true baseline. Urine specimens that were not initially spun down were sent to the University of Cincinnati for metallomics analysis by ICP-MS. Metallomic results were normalized to uCr levels.
All biological samples collected at WPAFB were processed and stored after each shift. While at JBC, samples were stored at -20°C until specimens were shipped back to the lab for processing.
Significance of demographic and lifestyle differences between bases and participant groups were analyzed using T-tests for continuous variables and using Fisher’s Exact tests for categorical variables due to small sample sizes. Pearson’s correlation calculations were used to assess relationships between symptom reports, range class characteristics, and instructor characteristics. An alpha of 0.05 was used to determine statistical significance. All survey data was analyzed using SAS software, Version 9.04.
Personal exposure monitoring results were reported as the mean ± the standard error of the mean (SEM). T-tests were used for statistical significance analysis between two groups. T-tests assuming unequal variance were used on data from real-time measurements attached to PEVs, significance was obtained when p < 0.05 unless indicated otherwise. For comparisons between multiple groups, a one-way analysis of variance (ANOVA) with Tukey’s post-hoc test was used.
Biological results were reported as the mean ± SEM. For comparisons among multiple groups, ANOVA with Tukey’s post-hoc test was used. For comparisons between groups, Student’s t-test was used for significance analysis. Pearson’s correlation calculations were used to identify correlations between measurements. Statistical significance was obtained when p < 0.05 unless indicated otherwise.