Abstract
Objective
To investigate the extent of fluid incursion throughout the respiratory tract of anesthetized pigs terminated using 3 different depopulation methods compared to termination by water submersion (H2O).
Methods
Inclusion criteria included pigs aged approximately 40 days. The study occurred over 2 consecutive days during spring. Pigs were anesthetized and terminated using their assigned method: water-based foam, high-expansion nitrogen foam, carbon dioxide gas, or H2O. Respiratory tracts were evaluated 3 ways: CT, gross examination, and histopathology. Immediately after death, pigs were CT scanned, and opacity changes were scored. After gross examination, representative tissues were scored microscopically.
Results
48 pigs were assigned to 4 treatment groups of 12 pigs each. Compared to all other methods, H2O had higher odds of increased opacities on CT in several proximal structures and bronchi and pulmonary parenchyma of multiple lung lobes. All groups had pulmonary petechiae to regional hemorrhages and subpleural hemorrhages. Emphysema aquosum was observed exclusively in H2O pigs (4 of 12 [33.3%]). Histologically, carbon dioxide gas had 97.8% decreased odds of pulmonary hemorrhage compared to H2O after accounting for sex.
Conclusions
Postmortem lesions and CT opacity patterns associated with water-based foam and high-expansion nitrogen foam are dissimilar to H2O. Foam-based methods involve mechanistic differences from drowning through either environmental or occlusional anoxia, not overt fluid inundation of airways.
Clinical Relevance
Foam-based methods are valuable candidates for US swine production and must be separated from the mislabel of drowning. Our findings provide new information regarding large-scale depopulation tools for emergency response efforts.
The US swine industry faces threats of disease outbreaks, introduction of foreign animal diseases, and natural disaster events. The large impact these situations could have on animal health and economics would require emergency herd depopulation to preclude animal morbidity and initiate immediate disease containment. Ideally, depopulation methods should satisfy certain criteria, such as rapid loss of animal consciousness and death, high throughput with short depopulation procedure turnaround times, protection of operating personnel (both mental and physical health), and accessibility/affordability of supplies/equipment. Despite this, several swine depopulation methods face notable shortcomings, such as low efficiency, high personnel fatigue, or supply inaccessibility that can occur, especially for methods that heavily rely on gas supply, such as CO2.1 Additionally, depopulation methods have varied effects on personnel distress, public perception, and staff burnout.2,3 Thus, selecting methods with reduced audiovisual sensory input and decreased contact with animals as they are depopulated may help mitigate such adverse effects.
Due to the limitations in current options for depopulation, there is a dire need for alternative, emerging technologies. Since 2006, class A water-based foam (WBF) has been a depopulation agent used by the USDA for floor-reared poultry.4 Water-based foam consists of a 1% solution of WBF concentrate in water, expanded using air to a ratio of 40 to 50:1. Investigations into extending the use of WBF to swine have repeatedly shown 100% mortality in pigs of various production phases, making WBF a strong candidate for swine depopulation.5–8 A promising alternative is high-expansion nitrogen (N2) foam (N2F), which comes in different forms depending on the instruments used for foam generation and the source of bubble disruption (ie, animal movement vs application of pulse). Generally, N2F composition differs from WBF in that less water is required for the foam mixture, which is injected with N2 gas to create large bubbles. Like WBF, N2F has shown success in depopulating poultry species,9 along with various ages and sizes of swine,10,11 and has been deemed suitable for use in Europe for stunning both poultry and swine.12 Studies10,11,13,14 have suggested that N2F causes death in pigs by creating anoxic conditions, but the underlying mechanism of death has not been defined as anoxia from breathing pure environmental N2 gas, mechanical occlusion of the proximal respiratory tract by the foam, or a combination of both. Similarly, while it has been speculated that WBF causes anoxia through airway occlusion, there are no comprehensive studies on this topic to date.
At this time, the effect of these foams on the respiratory system during the perimortem period is poorly understood. Given that they are a combination of liquid and gas components, we lack knowledge of whether these agents function more similarly to a gaseous or liquid compound once applied to pigs. Consequently, there has been ongoing speculation that using foam technology as a means of animal depopulation is equivalent to water submersion (H2O), or animal drowning. Specifically, drowning is a result of inundating the respiratory tract with fluid, resulting in respiratory system damage and dysfunction, hypoxemia, and subsequent organ death.15,16 On the contrary, the means by which foam-based technology functions is suggested to be occlusional, such as mechanically obstructing the nares and preventing the inspiration of atmospheric air, or by releasing gas (eg, N2 gas) and causing the depletion of environmental O2. Left unaddressed, foam-based methods are subject to being labeled as inhumane and “highly distressing,” restricting their use as a means of swine depopulation due to such misrepresentation.17 Therefore, our objectives were to (1) characterize postmortem changes in opacities within the airways using advanced imaging (CT); (2) describe observed tracheal inflammation, pulmonary hemorrhage, and any presence of emphysema aquosum within the respiratory tract; and (3) make comparisons of these changes between pigs that were terminated using WBF, N2F, carbon dioxide (CO2), and H2O. Our overall goal was to generate evidence to further inform swine industry stakeholders and regulatory officials on the decision-making process and selection of permissible depopulation methods in preparation for emergency situations involving swine herds. Our hypothesis was that pigs terminated using WBF, N2F, and CO2 would have decreased odds of fluid incursion, measured by changes in opacities and postmortem lesions, within the distal respiratory tract compared to pigs in the H2O group.
Methods
Animal enrollment and anesthesia
All animals included in this study were handled according to University of Minnesota IACUC protocol 2403-41911A and The Ohio State University protocol 2020A00000036. To determine the minimum number of pigs required for our study, a power analysis was performed. Assuming that 60% of pigs terminated using WBF will display foam penetration extending to the level of the tracheal bifurcation and 0% of pigs in the N2F group will display penetration to the tracheal bifurcation—with a confidence level of 95%, power of 0.9, 1:1 sample size ratio between groups, and a one-tailed test—we determined a sample size of 12 pigs/treatment.
A total of 48 pigs of mixed sex, averaging approximately 14 kg and 40 days of age, were enrolled in this study. All pigs originated from the same supplier and lacked clinical signs of respiratory disease (eg, sneezing, coughing, dyspnea, oculonasal discharge). Pigs were conveniently assigned to 4 treatment groups of 12 pigs each: WBF, N2F, CO2, and H2O. Each pig was anesthetized using injections of medetomidine/vatinoxan plus ketamine (10 mg/m2 + 10 mg/kg, IM). Pigs were monitored until a surgical plane of anesthesia was achieved. Additional medetomidine/vatinoxan plus ketamine (2 mg/m2 + 12 mg/kg) was administered if pigs did not display adequate sedation (eg, lack of eyelid, conjunctival, and limb reflexes) after 30 minutes of monitoring. This protocol was chosen to provide anesthesia without significant respiratory depression as determined by a pilot study performed in swine. The H2O pigs were anesthetized using tiletamine-zolazepam, ketamine, and xylazine (0.04 to 0.05 mL/kg) with additional xylazine (3.7 mg/kg) to ensure a deep plane of anesthesia (eg, lack of reflexes and complete relaxation of muscles) prior to submersion given the welfare implications of depopulating animals by drowning.17 Pigs were monitored until a surgical plane of anesthesia was achieved, which was confirmed based on the lack of response to noxious stimuli, lack of a corneal reflex, and maintenance of spontaneous ventilation. Once anesthetized, pigs were individually placed within method-specific depopulation receptacles and kept in sternal recumbency throughout the application of the assigned termination method.
Water-based foam generation and administration
Water-based foam was generated using a gas-powered water pump and a medium-expansion foam nozzle. A 1% WBF solution was pumped into a 50.2-m3 bulk tote bin (3.9 X 3.9 X 3.3 m, length X width X height) as previously outlined by Arruda et al.5 Individual anesthetized pigs were kept in the foaming receptacle for a total dwell time of 7.5 minutes based on previously established recommendations.5
Nitrogen foam generation and administration
A 0.73-m3 (1.2 X 0.80 X 0.76 m, length X width X height;) box made of high-density polyethylene resin plastic (HEFT AB) was connected to a cylinder of compressed N2 gas and a 21-L carboy of foam solution. A 5% concentration of rapeseed soap was created by mixing 1 L soap in 20 L water to make the foam solution. Rapeseed soap foam was used in this study as it was previously deemed compatible with and included with the N2F-generating system and is regularly available, plant based, and nontoxic. The anesthetized pig was placed in sternal recumbency within the box, which was then closed by placing a lid with a clear plexiglass panel for visualization on top of the box. The box was filled with approximately 15-mm-diameter foam bubbles made from mixing > 98% pure N2 gas and foam solution, and O2 was pushed out of a vent at the top of the box by delivering foam for 11 to 25 seconds. Once the box was full of bubbles, a pulse of N2 gas was applied to burst the bubbles. Oxygen levels were measured with sensors (O2 Sensor NUL-205 [#369634] with a Digital Display Unit [#369730] and Battery Pack Module [#369732]; NeuLog) at the N2 outlet and within the box, from which O2 levels were recorded as sequential readings. Anoxic conditions were confirmed once O2 levels in the box reached < 2% based on O2 readings from the sensors. Anoxia was held for 7.5 minutes.
CO2 gas administration
Anesthetized pigs were maintained in sternal recumbency with the aid of concrete blocks to support the pig upright inside of a 0.63-m3 (1.07 X 1.22 X 0.48 m, length X width X height) sealed, airtight CO2 chamber. The chamber was filled directly using CO2 cylinders. The gas flow was set at a 20% displacement rate by volume/min.2,18 Upon filling, pigs were left undisturbed for a dwell time of 15 minutes.
Water submersion
Anesthetized pigs were placed inside of a 0.59-m3 oval, open-top poly stock tank (0.9 X 1.3 X 0.5 m, length X width X height), which was filled with water to ensure full submersion of pigs. Pigs were manually maintained in sternal recumbency and submerged for a total of 7.5 minutes.
Confirmation of death
After the completion of each respective dwell time, death was confirmed by observing the absences of spontaneous respiration, corneal reflex, and auscultable heartbeat.
Computed tomography
Immediately after the confirmation of death, each carcass was removed and gently wiped dry using absorbent towels to remove any remaining residue (foam or water). Carcasses were transported to the CT scanning table in sternal recumbency on a padded sled, with care taken to ensure they remained in position to minimize the movement of any liquid or foam within the respiratory tract. Each pig was scanned immediately after placement on the table. Digital files were saved on disks for blinded interpretation by a single American College of Veterinary Radiology–boarded veterinary radiologist (EMG). Increased opacity on CT was considered suggestive of gas being replaced by a fluid, whether due to pulmonary edema, hemorrhage, or fluid incursion via aspiration or ingestion.19,20 The percentage of fluid opacity of 20 structures of the respiratory tract (Figure 1) postmortem was evaluated and assigned a score using a 5-point grading scale (0, no changes; 1, 1% to 25%; 2, 26% to 50%; 3, 51% to 75%; and 4, 76% to 100%; Figure 2). These structures included left and right nasal cavities, left and right frontal sinuses, left and right maxillary sinuses, pharynx, trachea, and both bronchi and pulmonary parenchyma for the left and right cranial, right middle, accessory, and left and right caudal lung lobes.
Diagram of the respiratory structures of interest during the CT evaluation of pigs. For each specified lung lobe, both the bronchus and pulmonary parenchyma were evaluated.
Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.25.01.0035
Representative CT slices of the swine respiratory tract demonstrating the scoring system. No pigs were assigned a score of 4 within the nasal cavity.
Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.25.01.0035
Postmortem evaluation
Immediately after each pig carcass was scanned, a postmortem examination was performed by a veterinary anatomic pathology resident (JYP), who was blinded to the depopulation method used on each carcass. Gross examination was focused on the respiratory tract, extending from the nares to the lungs. The lesions assessed included noting the gross extent of fluid penetration (unremarkable, proximal trachea, tracheal bifurcation, or terminal bronchi) along with the type and severity of pulmonary hemorrhages. Moreover, changes that suggested pathologic pulmonary overdistention, specifically rib impressions and emphysema aquosum, were noted to calculate the proportion of pigs affected per treatment group. Representative samples of the proximal trachea, tracheal bifurcation, and caudodorsal lung lobe were taken and immediately placed in 10% neutral-buffered formalin. If there were noticeable variations in a lesion (ie, pulmonary hemorrhage) throughout the caudodorsal lung fields, the most severely affected area was sampled. Lung samples were infused with formalin using an 18-gauge hypodermic needle and syringe to preserve airway expansion. For each pig, a section of proximal trachea, trachea at the level of the tracheal bifurcation, and lung were trimmed by JYP and processed by The Ohio State University’s Comparative Pathology & Digital Imaging Shared Resource. Tissues were embedded in paraffin wax, cut, and stained with H&E for interpretation. All slides were interpreted by JYP using a previously established rubric to score tracheitis (0, no significant microscopic changes; 1, lamina proprial inflammation; 2, superficial submucosal inflammation; 3, deep submucosal inflammation; and 4, transmural inflammation) and pulmonary hemorrhage (0, no changes; 1, 1% to 25%; 2, 26% to 50%; 3, 51% to 75%; and 4, 76% to 100% of parenchyma affected).21
Statistical analysis
A Spearman rank test was performed to assess the correlation between the CT scores for each structure within the respiratory tract and between measurements on the left or right side per sinus and lung lobe. Correlation coefficient (ρ) values of 0.1 to 0.3, 0.4 to 0.6, 0.7 to 0.9, and 1.0 were regarded as indicating weak, moderate, strong, and perfect correlation, respectively.22 Separate ordinal logistic regression models were constructed to investigate the effect of method (CO2, H2O, N2F, and WBF) on the severity of change in opacity (0, normal; 1, 1% to 25%; 2, 26% to 50%; 3, 51% to 75%; and 4, 76% to 100%) for each structure of interest within the respiratory system (left and right sinuses, left and right frontal sinuses, left and right maxillary sinuses, pharynx, trachea, left and right cranial bronchi, left and right caudal bronchi, right middle bronchus, accessory bronchus, accessory lung, left and right cranial lung, right middle lung, and left and right caudal lung). All models accounted for animal sex (male or female). An ordinal logistic regression model was used to also analyze the effect of method on the scored microscopic lesions of tracheitis (trachea, trachea at level of bifurcation) and pulmonary hemorrhage (caudodorsal lung). All comparisons were made with H2O set as the referent group. The results are described as either increased or decreased ORs of increased degrees of severity of the outcomes of opacities on CT, tracheitis, and pulmonary hemorrhage.
Due to the inherent lack of independence between structures within the respiratory system due to anatomical arrangement, such as physiological differences between lung lobes, a model for each outcome variable was created. If discrepancies in results between the left and right sides were observed, both results are presented for interpretation. Otherwise, model data from only 1 measurement is presented. Utilizing a Bonferroni correction23 for the 10 total models, a P value cutoff was determined as < .005. Tendencies corresponding to P values of .005 ≤ P < .010 are described.
Results
Confirmation of death
All 48 pigs enrolled in the study were confirmed dead after a single application of their assigned termination method. A secondary method of euthanasia was not used on any pig.
Confirmation of anoxic conditions for pigs terminated using N2F
For the N2F treatment, the onset of anoxic conditions (< 2% O2 in air) occurred within 90 seconds of treatment initiation for 11 of 12 pigs (range, 41 to 90 seconds), and hypoxic conditions (< 10% O2 in air) were achieved within 42 seconds for all pigs. For the single pig that was not exposed to anoxic conditions within 42 seconds, O2 levels were < 3.0% within 88 seconds and held below 3.0% for 389 seconds. The average amount of time N2F treatment pigs were in anoxic conditions was 360 seconds (range, 31 to 474 seconds). The average amount of time in hypoxic conditions with O2 levels between 2.0 and 9.9 was 448 seconds (range, 434 to 454 seconds; Supplementary Table S1). Anoxic conditions were achieved for more than 90% of pigs, and 100% of the pigs were held in conditions with substantially low O2 levels for durations of time sufficient to cause death.
Computed tomography
The Spearman rank correlation showed strong correlations between multiple structures that were evaluated on CT given the inherent anatomical order of these structures within the respiratory tract. These included a strong correlation between left and right nasal (ρ = 0.99) and maxillary (ρ = 0.72) sinuses, a moderate correlation between the trachea and pharynx (ρ = 0.66), and moderate-to-strong correlations within the bronchial (ρ range, 0.61 to 0.89) and lung measurements (ρ range, 0.45 to 0.9) within their respective regions. Weak-to-moderate correlations were found between maxillary sinuses and the frontal or nasal sinuses (ρ range, 0.12 to 0.59). Mainly weak or negative correlations were found for the maxillary, frontal, or nasal sinuses (ρ range, −0.22 to 0.29) and lower structures of the respiratory system. A moderate correlation was observed between the maxillary sinuses and the accessory lung (ρ = 0.37), trachea, and right frontal sinus (ρ = 0.32).
The main method-associated effect on upper respiratory structures was observed for the right maxillary sinus, where the odds of increased opacity tended to be 87% lower for CO2 (OR, 0.13; P = .01) compared to H2O (Table 1). For the pharynx, the odds for having increased opacity were 94%, < 99.9%, and 92% lower for CO2 (OR, 0.06; P = .002), N2F (OR, 0.001; P < .001), and WBF (OR, 0.08; P = .004), respectively, compared to H2O. Within the bronchial tree, a similar pattern was observed for the cranial and caudal bronchi, where CO2, N2F, and WBF had lower odds of increased opacities compared to H2O (P < .01).
Results for significant ordinal logistic regression models on the odds of the presence of changes in opacity on CT after depopulation using 4 different methods.
| Variable | Category | OR | SE | 95% CI | P value | |
|---|---|---|---|---|---|---|
| Pharynx | Methodology | H2O | Referent | |||
| CO2 | 0.06 | 0.05 | 0.009–0.34 | .002 | ||
| N2F | 0.001 | 0.01 | 0.0007–0.02 | < .001 | ||
| WBF | 0.08 | 0.07 | 0.015–0.46 | .004 | ||
| Left cranial bronchus | Methodology | H2O | Referent | |||
| CO2 | 0.01 | 0.01 | 0.001–0.11 | < .001 | ||
| N2F | 0.003 | 0.004 | 0.0002–0.04 | < .001 | ||
| WBF | 0.02 | 0.02 | 0.003–0.15 | < .001 | ||
| Right caudal bronchus | Methodology | H2O | Referent | |||
| CO2 | 0.002 | 0.003 | 0.0001–0.03 | < .001 | ||
| N2F | N/Aa | N/Aa | N/Aa | .991 | ||
| WBF | 0.02 | 0.03 | 0.002–0.27 | .002 | ||
| Accessory lung | Methodology | H2O | Referent | |||
| CO2 | 0.07 | 0.06 | 0.01–0.39 | .003 | ||
| N2F | 0.01 | 0.01 | 0.001–0.11 | < .001 | ||
| WBF | 0.10 | 0.08 | 0.02–0.52 | .006 | ||
| Left caudal lung | Methodology | H2O | Referent | |||
| CO2 | 0.63 | 0.47 | 0.14–2.70 | .53 | ||
| N2F | 0.08 | 0.07 | 0.16–0.41 | .002 | ||
| WBF | 0.08 | 0.07 | 0.015–0.41 | .003 | ||
| Right cranial lung | Methodology | H2O | Referent | |||
| CO2 | 0.07 | 0.06 | 0.01–0.37 | .002 | ||
| N2F | 0.05 | 0.04 | 0.01–0.27 | .001 | ||
| WBF | 0.11 | 0.01 | 0.02–0.58 | .009 | ||
| Right middle lung | Methodology | H2O | Referent | |||
| CO2 | 0.11 | 0.09 | 0.02–0.54 | .007 | ||
| N2F | 0.09 | 0.07 | 0.02–0.42 | .002 | ||
| WBF | 0.06 | 0.05 | 0.01–0.33 | .001 | ||
CO2 = Carbon dioxide gas. H2O = Water submersion. N2F = Nitrogen foam. N/A = Not applicable. WBF = Water-based foam.
All 12 observations were completely determined and thus could not be estimated by the model.
Opacity scores obtained from CT imaging, per structure and termination method, are further outlined in Table 2.
Descriptive statistics of CT image opacity scores (0, normal; 1, 1% to 25%; 2, 26% to 50%; 3, 51% to 75%; and 4, 76% to 100%) for 48 pigs (12 pigs/method) for each method for significant ordinal logistic regression models.
| Opacity score | ||||||
|---|---|---|---|---|---|---|
| Method | Normal | 1–25% | 26–50% | 51–75% | 75–100% | |
| 0 | 1 | 2 | 3 | 4 | ||
| Left maxillary sinus | H2O | 5a (41.7)b | 4 (33.3) | 1 (8.3) | 1 (8.3) | 1 (8.3) |
| CO2 | 3 (25.0) | 1 (8.3) | 0 (0) | 5 (41.7) | 3 (25.0) | |
| N2F | 2 (16.7) | 2 (16.7) | 4 (33.3) | 1 (8.3) | 3 (25.0) | |
| WBF | 7 (58.3) | 0 (0) | 2 (16.7) | 1 (8.3) | 2 (16.7) | |
| Total | 17 (35.4) | 7 (14.6) | 7 (14.6) | 8 (16.7) | 9 (18.8) | |
| Right maxillary sinus | H2O | 6 (50.0) | 3 (25.0) | 1 (8.3) | 0 (0) | 2 (16.7) |
| CO2 | 2 (16.7) | 0 (0) | 0 (0) | 7 (58.3) | 3 (25.0) | |
| N2F | 1 (8.3) | 2 (16.7) | 4 (33.3) | 3 (25.0) | 2 (16.7) | |
| WBF | 5 (41.7) | 2 (16.7) | 0 (0) | 2 (16.7) | 3 (25.0) | |
| Total | 14 (29.2) | 7 (14.6) | 5 (10.4) | 12 (25.0) | 10 (20.8) | |
| Pharynx | H2O | 0 (0) | 3 (25.0) | 4 (33.3) | 3 (25.0) | 2 (16.7) |
| CO2 | 0 (0) | 0 (0) | 1 (8.3) | 3 (25.0) | 8 (66.7) | |
| N2F | 2 (16.7) | 9 (75.0) | 1 (8.3) | 0 (0) | 0 (0) | |
| WBF | 0 (0) | 1 (8.3) | 5 (41.7) | 5 (41.7) | 1 (8.3) | |
| Total | 2 (4.2) | 13 (27.1) | 11 (22.9) | 11 (22.9) | 11 (22.9) | |
| Left cranial bronchus | H2O | 9 (75.0) | 3 (25.0) | 0 (0) | 0 (0) | 0 (0) |
| CO2 | 0 (0) | 2 (16.7) | 0 (0) | 0 (0) | 10 (83.3) | |
| N2F | 11 (91.7) | 1 (8.3) | 0 (0) | 0 (0) | 0 (0) | |
| WBF | 7 (58.3) | 3 (25.0) | 2 (16.7) | 0 (0) | 0 (0) | |
| Total | 27 (56.3) | 9 (18.8) | 2 (4.2) | 0 (0) | 10 (20.1) | |
| Right caudal bronchus | H2O | 9 (75.0) | 3 (25.0) | 0 (0) | 0 (0) | 0 (0) |
| CO2 | 1 (8.3) | 0 (0) | 3 (25.0) | 1 (8.3) | 7 (58.3) | |
| N2F | 12 (100.0) | 0 (0) | 0 (0) | 0 (0) | 0 (0) | |
| WBF | 2 (16.7) | 6 (50.0) | 4 (33.3) | 0 (0) | 0 (0) | |
| Total | 24 (50.0) | 9 (18.8) | 7 (14.6) | 1 (2.1) | 7 (14.6) | |
| Right middle bronchus | H2O | 7 (58.3) | 3 (25.0) | 2 (16.7) | 0 (0) | 0 (0) |
| CO2 | 0 (0) | 0 (0) | 0 (0) | 1 (8.3) | 11 (91.7) | |
| N2F | 11 (91.7) | 6 (50.0) | 0 (0) | 0 (0) | 0 (0) | |
| WBF | 6 (50.0) | 6 (50.0) | 0 (0) | 0 (0) | 0 (0) | |
| Total | 24 (50.0) | 10 (20.8) | 2 (4.2) | 1 (2.1) | 11 (22.9) | |
| Accessory lung | H2O | 0 (0) | 8 (66.7) | 2 (16.7) | 0 (0) | 2 (16.7) |
| CO2 | 0 (0) | 0 (0) | 5 (41.7) | 2 (16.7) | 5 (41.7) | |
| N2F | 1 (8.3) | 10 (88.3) | 1 (8.3) | 0 (0) | 0 (0) | |
| WBF | 1 (8.3) | 5 (41.7) | 4 (33.3) | 1 (8.3) | 1 (8.3) | |
| Total | 2 (4.2) | 23 (47.9) | 12 (25.0) | 3 (6.3) | 8 (16.7) | |
| Left caudal lung | H2O | 0 (0) | 0 (0) | 4 (33.3) | 3 (25.0) | 5 (41.7) |
| CO2 | 0 (0) | 0 (0) | 2 (16.7) | 4 (33.3) | 6 (50.0) | |
| N2F | 0 (0) | 4 (33.3) | 5 (41.7) | 2 (16.7) | 1 (8.3) | |
| WBF | 1 (8.3) | 3 (25.0) | 5 (41.7) | 1 (8.3) | 2 (16.7) | |
| Total | 1 (2.1) | 7 (14.6) | 16 (33.3) | 10 (20.8) | 14 (29.2) | |
| Right cranial lung | H2O | 0 (0) | 8 (66.7) | 1 (8.3) | 1 (8.3) | 2 (16.7) |
| CO2 | 0 (0) | 0 (0) | 6 (50.0) | 2 (16.7) | 4 (33.3) | |
| N2F | 1 (8.3) | 8 (66.7) | 3 (25.0) | 0 (0) | 0 (0) | |
| WBF | 0 (0) | 6 (50.0) | 4 (33.3) | 1 (8.3) | 1 (8.3) | |
| Total | 1 (2.1) | 22 (45.8) | 14 (29.2) | 4 (8.3) | 7 (14.6) | |
| Right middle lung | H2O | 0 (0) | 7 (58.3) | 1 (8.3) | 3 (25.0) | 1 (8.3) |
| CO2 | 0 (0) | 0 (0) | 4 (33.3) | 3 (25.0) | 5 (41.7) | |
| N2F | 0 (0) | 6 (50.0) | 5 (41.7) | 1 (8.3) | 0 (0) | |
| WBF | 1 (2.1) | 7 (58.3) | 1 (8.3) | 2 (16.7) | 1 (8.3) | |
| Total | 1 (2.1) | 20 (41.7) | 11 (22.9) | 9 (18.8) | 7 (14.6) | |
Within the accessory lung lobe, compared to H2O, the odds of increased opacity were 93%, 99% and 90% lower for CO2 (OR, 0.07; P = .003), N2F (OR, 0.01; P < .001), and WBF (OR, 0.10; P = .006), respectively. Similarly, the odds of opacity changes in the right cranial and right middle lung lobes decreased between 90% to 95% for CO2, N2F, and WBF compared to H2O (P < .01). In the left caudal lung, however, CO2 did not differ from H2O (OR, 0.63; P = .53), whereas N2F (OR, 0.08; P = .002) and WBF (OR, 0.08; P = .003) had decreased odds of opacity changes. No effect of methodology was observed for the left and right nasal sinuses, frontal sinuses, accessory bronchus, trachea, or left caudal bronchus. Final model results with ORs, SEs, 95% CIs, and corresponding P values are presented in Table 1.
Gross and microscopic examination
The predominant lesion, both grossly and microscopically, was pulmonary hemorrhage of varying severity. This ranged from mild petechiae to marked regional hemorrhages along with occasional subpleural hemorrhages. A pulmonary hemorrhage was present in the majority (> 50%) of pigs regardless of the depopulation method used. Lymphoplasmacytic tracheitis, characterized by infiltrating aggregates of inflammatory lymphocytes and plasma cells, was noted in the proximal trachea and tracheal bifurcation. Across all treatment groups, the degree of tracheitis, when present, varied from limited to the superficial lamina propria to transmural (full thickness). The tracheitis scores were not statistically different across treatment groups. Microscopically, the only significant difference in pulmonary hemorrhage between treatment groups was the 97.8% decrease in odds of pulmonary hemorrhage due to CO2 compared to H2O after accounting for sex (95% CI, 0.003 to 0.170; P < .001; Table 3).
Mean ± SD scores of histologic lesions seen in terminated pigs.
| Tracheitis (proximal trachea) | Tracheitis (tracheal bifurcation) | Pulmonary hemorrhage | |
|---|---|---|---|
| CO2 | 2.08 ± 0.67 | 2.33 ± 0.98 | 0.92 ± 1.31 |
| N2F | 2.17 ± 0.72 | 2.25 ± 0.87 | 1.83 ± 0.94 |
| WBF | 1.67 ± 0.49 | 1.75 ± 0.62 | 2.75 ± 1.14 |
| H2O | 1.92 ± 0.67 | 2.25 ± 1.22 | 2.92 ± 1.31 |
Grossly, only pigs in the H2O group had pulmonary parenchymal lesions of prominent interlobular septae due to fluid overdistention and exuded a colorless, translucent fluid resembling water on cross-sectioning. This colorless fluid also often accumulated within the subpleural space (Figure 3). This lesion, called “emphysema aquosum,” was present in 4 of 12 of the H2O pigs (33%). The lungs sampled from the H2O pigs exhibiting emphysema aquosum were markedly heavy and sank in formalin upon sampling. A similar colorless fluid extended into the distal trachea at the level of the tracheal bifurcation. Rib impressions were observed on the lungs of 4 of 12 H2O pigs only (33%). Additionally, a foamy, light-tan to pink-tinged fluid was observed within the respiratory tracts of a total of 28 of 48 study pigs (58.3%). This described fluid was present in at least 1 pig of all treatment groups. The appearance of this tan/pink fluid most closely resembled pulmonary edema and could not conclusively be visually discerned from the foams used for depopulation. The gross extent of tan/pink fluid within the respiratory tract was 9 of 12 (75%) in the distal principal bronchi and 3 of 12 (25%) grossly inapparent for H2O pigs; 1 of 12 (8.3%) in the tracheal lumen, tracheal bifurcation, and distal principal bronchi, respectively, and 9 of 12 (75%) grossly inapparent for N2F pigs; and 5 of 12 (41.7%) in the tracheal lumen, 2 of 12 (16.7%) in the tracheal bifurcation, 1 of 12 (8.3%) in the distal principal bronchi, and 4 of 12 (33.3%) grossly inapparent for WBF pigs. Fluid was grossly inapparent for 12 of 12 pigs in the CO2 group (100%).
Example of emphysema aquosum observed in the water submersion group. A—Interlobular septae (fine arrows) are prominent due to fluid overdistension. Rib impressions (asterisks) were seen due to compression of the lungs against the ribcage. B—Subpleural accumulation of water (broad arrow), which exuded from lungs on cross-sectioning. Asterisks indicate rib impressions due to overdistention of the lungs and compression against the ribcage.
Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.25.01.0035
Discussion
In this study, advanced imaging using CT and postmortem evaluation revealed that lesions in pigs terminated using WBF and N2F depopulation methods are unlike those found in pigs terminated using H2O, or drowning. Implementing highly detailed serial imaging immediately postmortem, along with gross and microscopic examinations for respiratory pathology, allowed us to identify distinct patterns between treatment groups, particularly between the 2 foam-based methods and H2O. We included real-time advanced imaging (CT) in our study in addition to gross and microscopic examinations to provide greater detail in our observations. Computed tomography is an imaging modality that is commonly used in veterinary medicine given its benefits of serial images, notably shorter scan times, and soft tissue contrast that is appropriate for respiratory tract evaluation.24 Thus, CT imaging of whole carcasses provided us the ability to obtain data without dislodging fluids and foams, which could otherwise result in artifactual changes in opacity scores and distributions.
At this time, we are not aware of other studies that utilize CT imaging as part of investigating depopulation methods’ effects when applied to animals. Our goal was to provide a multifaceted approach to determining how WBF and N2F travel within the respiratory tract and thus how they may function. Previously, Lorbach et al8 noted that 7 of 12 sows (58.3%) terminated using compressed N2 foam and 10 of 12 (83.3%) terminated with aspirated foam had foam grossly visible at or beyond the level of the tracheal bifurcation. Considering that the majority (> 50%) of sows showed penetration to these structures, whereas our study found that only 16.6% of N2F and 25.0% of WBF pigs had fluid grossly within this region, this could suggest respiratory tract involvement greater than that shown in our study’s gross findings. However, we suspect that these discrepancies may be from differences in study design, with our study involving nursery pigs (vs cull sows) along with our priority in maintaining animal positioning and measuring data as close to animal termination as possible, thus minimizing artifactual fluid displacement. However, it will be interesting to perform a similar study as this using larger swine to identify whether foam-based technologies could function differently based on pig size.
Interestingly, CT evaluation of CO2 and both foam-based methods revealed relative sparing of the bronchial tree, of the majority of the pulmonary parenchyma, and proximally in multiple sinuses on CT, whereas a higher degree of opacities was found for H2O in these areas.
On gross evaluation, one critical finding unique to the H2O pigs was prominent interlobular septae and pulmonary overdistention with colorless fluid resembling water. This lesion is highly consistent with a characteristic phenomenon called “emphysema aquosum,” which is used to identify fatal drowning in both human and veterinary forensic cases.25–29 Emphysema aquosum occurs due to the inhalation of water in addition to air, manifesting as heavy, “water-logged” lungs.25–27 Additionally, the greater severity of gross and microscopic pulmonary hemorrhages observed in H2O pigs, in comparison to all other methods, suggests that H2O submersion likely results in more extensive damage of the pulmonary vasculature and parenchyma, affecting the ability to ventilate and perform gas exchange (respiration) at the cellular level.15
The limited extent of CT opacities and the absence of emphysema aquosum in both foam-based methods overall suggest that the mixture of gases (air or N2) with liquid components creates a medium that functions differently from overt fluid inundation of the airways (ie, drowning). For WBF, CT opacities were heavily limited to around the uppermost airway structures, potentially affecting the inspirational phase of ventilation. Specifically, they likely cause mechanical obstruction at the level of within the nares or the level of the upper sinuses. For N2F, the paucity of opacities in the proximal and deeper structures supports that, as it does in poultry,9,13 N2F creates environmental oxygen depletion, by definition anoxia, by displacing oxygen out of the containers, releasing N2 gas as gas-filled foam bubbles are broken down.13,30
While CT imaging provides many advantages for the goals of our study, there are some limitations. On imaging alone, increased opacities within the respiratory tract (interpreted as fluid)20,24,25 cannot distinguish between the nature of those fluids. While these changes certainly could be from an exogenous source (ie, drowning), other differentials are pulmonary hemorrhage and pulmonary edema, both of which were changes seen during postmortem evaluation. Pulmonary hemorrhage and edema are common agonal changes and can occur during perimortem asphyxia in both humans and veterinary species.28,31,32 Human autopsies documented from 500 human drowning events indicated a majority of cases (61%) involving pulmonary edema,33 which was similarly noted in drowned dogs.34,35 Noncardiogenic pulmonary edema, which characteristically manifests in the caudodorsal lung fields, can result from perimortem acute respiratory distress syndrome, which may occur during drowning events.36 This could partially explain the lack of CT differences between CO2 and H2O in the left caudal lung lobe along with the lack of differences across groups in the trachea, which is a location that often contains pulmonary edema by agonal reflux from the lungs.
Finally, while great efforts were made to minimize movement during and after the termination of each pig, we must consider the possibility that any movement, such as from wiping foamy residue off of pigs, could have dislodged fluids and affected the CT results of some pigs. Another consideration is that method-specific times to animal death may affect the duration of respiratory distress in individual pigs. Given that the length of time to pig death by drowning has not been previously investigated, prolongation of the perimortem period, resulting in greater amounts of fluid inhalation and movement, could have increased the degree of opacities in H2O pigs in comparison to the other methods. Therefore, comparisons that account for these times are needed to better interpret the degrees of these relative differences.
Ultimately, this study highlights that there are stark differences between pure gas and foam-based swine depopulation from H2O submersion, and it addresses the concern of whether foam-based methods are mechanistically similar to drowning. The results elucidate the mechanisms by which WBF and N2F function, with the former causing asphyxia by proximal airway occlusion and the latter involving environmental O2 depletion, especially for the N2F method, as supported by data from the oxygen sensors (Supplementary Table S1). Our methods provide objective data that can help mitigate negative perceptions and welfare concerns by preventing foam-based methods from being mislabeled as animal drowning.
Working toward the understanding and availability of foam-based technology would facilitate disease outbreak and emergency response efforts by providing large-scale depopulation tools for US swine producers. While N2F is successful in inducing a rapid loss of consciousness and death in poultry and swine,10,14,30 swine industry stakeholders have expressed concerns regarding access to N2 supplies in urgent situations along with the cost and labor of the specialized N2F equipment.37 Similarly, stakeholders perceived that sudden increases in CO2 demand could easily cause supply shortages,1,37 limiting the ability to execute CO2 depopulation at commercial herd volumes. The efficacy and relative ease of use of WBF aims to mitigate these concerns by catering to existing equipment and foam supplies in the National Veterinary Stockpile as well as off-the-shelf components that are easily accessible. Thus, we can ultimately work toward counteracting the potential logistical bottlenecks for specialized labor or equipment in times of emergency situations.
Supplementary Materials
Supplementary materials are posted online at the journal website: avmajournals.avma.org.
Acknowledgments
The authors would like to thank Dawn Torrisi, Hannah Cochran, Natalie Tarbuck, Madison Owsiany, Erika Stevens, and Marissa Hall for their involvement throughout the research trials. They also thank The Ohio State University’s Comparative Pathology & Digital Imaging Shared Resource for their work in processing the histology slides for our study, Ms. Margret Tuisalo’o work at the University of Minnesota, and HEFT AB for leasing the equipment and supplies for N2F methods.
Disclosures
Dr. Aarnes is a member of the JAVMA Scientific Review Board, but was not involved in the editorial evaluation of or decision to accept this article for publication.
No AI-assisted technologies were used in the composition of this manuscript.
Funding
This research was funded by National Pork Checkoff (contract No. PR-005723).
ORCID
Janice Y. Park https://orcid.org/0000-0003-4226-7339
Andréia G. Arruda https://orcid.org/0000-0001-7040-9686
Andrew S. Bowman https://orcid.org/0000-0002-0738-8453
References
- 1.↑
CO2 shortage in 2022: a threat to food, beverage, and healthcare. Everstream Analytics. Accessed February 10, 2025. https://www.everstream.ai/risk-centers/co2-shortage-2022-industry-threat/
- 2.↑
Baysinger A, Kogan LR. Mental health impact of mass depopulation of swine on veterinarians during COVID-19 infrastructure breakdown. Front Vet Sci. 2022;9:842585. doi:10.3389/fvets.2022.842585
- 3.↑
Purc-Stephenson RJ, Doctor J. “The hill in front of you”: a qualitative study of the mental health impact of livestock diseases and depopulation on farmers. J Agromedicine. Published online February 22, 2025. doi:10.1080/1059924X.2025.2470967
- 4.↑
Benson ER, Alphin RL, Rankin MK, Caputo MP, Johnson AL. Electroencephalogram-based methodology for determining unconsciousness during depopulation. Avian Dis. 2012;56(4):884–890. doi:10.1637/10163-040912-Reg.1
- 5.↑
Arruda AG, Campler MR, Cheng T, et al. Reliability of water-based medium-expansion foam as a depopulation method for nursery pigs and cull sows. Transbound Emerg Dis. 2022;69(5):e2719–e2730. doi:10.1111/tbed.14622
- 6.
Campler MR, Cheng TY, Arruda AG, et al. Refinement of water-based foam depopulation procedures for finisher pigs during field conditions: welfare implications and logistical aspects. Prev Vet Med. 2023;217:105974. doi:10.1016/j.prevetmed.2023.105974
- 7.
Korenyi-Both J, Vidaurre J, Held T, et al. Description of electroencephalographic data gathered using water-based medium-expansion foam as a depopulation method for nursery pigs. Sci Rep. 2022;12(1):16798. doi:10.1038/s41598-022-21353-7
- 8.↑
Lorbach JN, Campler MR, Youngblood B, et al. Comparison of gaseous and water-based medium-expansion foam depopulation methods in cull sows. Animals (Basel). 2021;11(11):3179. doi:10.3390/ani11113179
- 9.↑
Gurung S, White D, Archer G, et al. Carbon dioxide and nitrogen infused compressed air foam for depopulation of caged laying hens. Animals (Basel). 2018;8(1):6. doi:10.3390/ani8010006
- 10.↑
Validation and demonstration of utilizing high expansion nitrogen foam for large scale depopulation of swine. Pork Checkoff. June 23, 2022. Accessed January 5, 2025. https://porkcheckoff.org/research/validation-and-demonstration-of-utilizing-high-expansion-nitrogen-foam-for-large-scale-depopulation-of-swine/
- 11.↑
Lindahl C, Sindhøj E, Brattlund Hellgren R, Berg C, Wallenbeck A. Responses of pigs to stunning with nitrogen filled high-expansion foam. Animals (Basel). 2020;10(12):2210. doi:10.3390/ani10122210
- 12.↑
Nielsen SS, Alvarez J, Bicout DJ, et al. The use of high expansion foam for stunning and killing pigs and poultry. EFSA J. 2024;22(7):e8855. doi:10.2903/j.efsa.2024.8855
- 13.↑
McKeegan DEF, Reimert HGM, Hindle VA, et al. Physiological and behavioral responses of poultry exposed to gas-filled high expansion foam. Poult Sci. 2013;92(5):1145–1154. doi:10.3382/ps.2012-02587
- 14.↑
Towers L. Humane euthanising using nitrogen foam. The Pig Site. January 3, 2013. Accessed January 17, 2025. https://www.thepigsite.com/news/2013/01/humane-euthanising-using-nitrogen-foam-1
- 15.↑
McCall JD, Sternard BT. Drowning: Clinical Management. In: StatPearls (Internet) Treasure Island. StatPearls Publishing; update published online August 8, 2023. Accessed February 26, 2025. http://www.ncbi.nlm.nih.gov/books/NBK430833/
- 16.↑
Frankel LR. Drowning and near drowning (submersion injuries) (chapter 356). American Academy of Pediatrics. Accessed February 26, 2025. https://publications.aap.org/pediatriccare/book/348/chapter/5788647/drowning-and-near-drowning-submersion-injuries
- 17.↑
Leary S, Pharmaceuticals F, Anthony R, et al. AVMA Guidelines for the depopulation of animals: 2019 edition. Published online 2019. Accessed January 5, 2025. https://www.avma.org/sites/default/files/resources/AVMA-Guidelines-for-the-Depopulation-of-Animals.pdf
- 18.↑
Rice M, Baird C, Stikeleather L, Morrow WEM, Meyer R. Carbon dioxide system for on-farm euthanasia of pigs in small groups. J Swine Health Prod. 2014;22(5):248–254. doi:10.54846/jshap/834
- 19.↑
Wang Z, Ma K, Zou D, et al. Diagnosis of drowning using postmortem computed tomography combined with endoscopic autopsy: a case report. Medicine (Baltimore). 2020;99(11):e19182. doi:10.1097/MD.0000000000019182
- 20.↑
Christe A, Aghayev E, Jackowski C, Thali MJ, Vock P. Drowning–post-mortem imaging findings by computed tomography. Eur Radiol. 2008;18(2):283–290. doi:10.1007/s00330-007-0745-4
- 21.↑
Torrisi DM, Campler MR, Park JY, et al. Evaluation of commercially available class A water-based foam concentrates for swine depopulation. Master’s thesis. The Ohio State University; 2024. Accessed January 16, 2025. http://rave.ohiolink.edu/etdc/view?acc_num=osu1732629329091973
- 22.↑
Dancey CP, Reidy J. Statistics Without Maths for Psychology: Using SPSS for Windows. Prentice Hall; 2004.
- 23.↑
Bland JM, Altman DG. Multiple significance tests: the Bonferroni method. BMJ. 1995;310(6973):170. doi:10.1136/bmj.310.6973.170
- 24.↑
Greco A, Meomartino L, Gnudi G, Brunetti A, Di Giancamillo M. Imaging techniques in veterinary medicine. Part II: computed tomography, magnetic resonance imaging, nuclear medicine. Eur J Radiol Open. 2022;10:100467. doi:10.1016/j.ejro.2022.100467
- 25.↑
Jian J, Wan L, Shao Y, et al. Postmortem chest computed tomography for the diagnosis of drowning: a feasibility study. Forensic Sci Res. 2019;6(2):152–158. doi:10.1080/20961790.2018.1557386
- 26.
Tyr A, Heldring N, Winskog C, Zilg B. Diagnosing fatal drownings: a review of the postmortem findings. Forensic Sci Int. 2024;364:112251. doi:10.1016/j.forsciint.2024.112251
- 27.↑
Schneppe S, Dokter M, Bockholdt B. Macromorphological findings in cases of death in water: a critical view on “drowning signs”. Int J Legal Med. 2021;135(1):281–291. doi:10.1007/s00414-020-02469-9
- 28.↑
McEwen BJ. Nondrowning asphyxia in veterinary forensic pathology: suffocation, strangulation, and mechanical asphyxia. Vet Pathol. 2016;53(5):1037–1048. doi:10.1177/0300985816643370
- 29.↑
McEwen BJ, Gerdin J. Veterinary forensic pathology: drowning and bodies recovered from water. Vet Pathol. 2016;53(5):1049–1056. doi:10.1177/0300985815625757
- 30.↑
Culhane M. Advances in stunning methods – high expansion nitrogen foam and its potential uses in swine and poultry. Presented at: Animal Care and Handling Conference of the North American Meat Institute; May 25, 2023; Kansas City, Missouri.
- 31.↑
Greaves P. Respiratory tract. In: Greaves P, ed. Histopathology of Preclinical Toxicity Studies. 4th ed. Academic Press; 2012:207–261. doi:10.1016/B978-0-444-53856-7.00006-3
- 32.↑
Swann HE. The development of pulmonary edema during the agonal period of sudden asphyxial deaths. J Forensic Sci. 1964;9(3):360–373.
- 33.↑
Girela-López E, Beltran-Aroca CM, Dye A, Gill JR. Epidemiology and autopsy findings of 500 drowning deaths. Forensic Sci Int. 2022;330:111137. doi:10.1016/j.forsciint.2021.111137
- 34.↑
Conn AW, Miyasaka K, Katayama M, et al. A canine study of cold water drowning in fresh versus salt water. Crit Care Med. 1995;23(12):2029–2037. doi:10.1097/00003246-199512000-00012
- 35.↑
Piegari G, De Biase D, d’Aquino I, et al. Diagnosis of drowning and the value of the diatom test in veterinary forensic pathology. Front Vet Sci. 2019;6:404. doi:10.3389/fvets.2019.00404
- 36.↑
Unger K, Martin LG. Noncardiogenic pulmonary edema in small animals. J Vet Emerg Crit Care (San Antonio). 2023;33(2):156–172. doi:10.1111/vec.13278
- 37.↑
Park JY, Campler MR, Cheng TY, et al. Perceptions of swine industry stakeholders on the use of water-based foam, high-expansion nitrogen foam, and carbon dioxide gas as methods of swine depopulation. Prev Vet Med. 2024;233:106356. doi:10.1016/j.prevetmed.2024.106356
