The Evaluation of the Containment Efficacy of Semi-Rigid Isolators for Housing Cages of Laboratory Animals Infected With BSL-3 Agents

Abstract

Research animals models infected with Biosafety Level-3 (BSL-3) agents need to be housed in specialized biocontainment caging. Most of these specialized cages have input and exhaust that is high efficiency particulate air filtered and sealed to prevent escape of the BSL-3 agent. An alternative to the use of the above BSL-3 biocontainment caging is the use of a flexible film or modified semi-rigid plastic film isolator that has its own high efficiency particulate air–filtered input and exhaust and is sealed with respect to the animal room environment, thus preventing BSL-3 agent escape. Standard caging can be housed within such an isolator. Computational fluid dynamics was used to evaluate the integrity of modified semi-rigid isolators for containment of aerosolized BSL-3 agents. Three isolators were located inside an animal BSL-3 room to provide an extra tier of protection and to permit different infectious studies within the same room while reducing or eliminating the risk of cross-contamination. The isolators were sized to house caging for rabbits and smaller non-human primates such as marmosets, African greens, and macaques. Multiple case studies of failure scenarios were investigated, including isolator breaches through the plastic membrane seam separation and rips, and exhaust fan failure. Breaching the level of containment provided by the isolators required the improbable simultaneous event of a plastic membrane rip in addition to the rare malfunction of the back-up exhaust fans. Each isolator was equipped with 2 blower motors connected in parallel to a common exhaust plenum and a battery backup. Even with this rare double (simultaneous) event, the animal BSL-3 room air exhaust system was able to contain the few droplets released in the simulated computational fluid dynamics breach. The modified semi-rigid isolators with negative airflow proved safe and effective for aerosol studies using BSL-3 agents, even in the unlikely event of a breach in containment.

BIOCONTAINMENT HOUSING FOR ABSL-3

Research animal models infected with BSL- 3 agents need to be housed in specialized biocontainment caging. Specialized racks of multiple individually ventilated cages are available for smaller animals such as mice, rats, and other small rodents such as hamsters. Larger laboratory animals such as rabbits, ferrets, smaller New or Old World non-human primates (NHPs) such as marmosets or even younger macaques require somewhat larger custom-designed housing made of stainless-steel or aluminum frames. Each caging unit can be sealed when closed to prevent escape of the BSL-3 agent. Air is pulled from the Animal Biosafety Level 3 (ABSL-3) room to a high efficiency particulate air (HEPA)-filtered in-going air port and through the cage to an outgoing HEPA-filtered exhaust, creating a relatively negative pressure inside the biocontainment housing unit. There may be prefilters prior to the HEPA filter and a magnehelic gauge to assure a proper pressure differential is maintained. The exhaust air can then go back into the room or out through a facility exhaust. A further safety precaution for personal respiratory and ocular membrane protection is the use of a powered air purifying respirator, which helps to provide respiratory, head, face, and eye protection. Any animal removed from an ABSL-3 biocontainment cage to be transported would need to be placed into a specialized biocontainment transport cage/box with the same sealing requirements to prevent escape of the BSL-3 agent.

Also, any handling or procedures conducted on the BSL-3–infected animal must be carried out in a certified class II bio-safety cabinet (BSC) with each of the 2 laboratory personnel wearing 2 pairs of gloves, the outer pair removed inside the BSC after the procedure so that the remaining gloves are not contaminated, and the animal is returned to its enclosure.^1–3

SEMI-RIGID PLASTIC FILM ISOLATORS

An alternative to the use of ABSL-3 biocontainment cages is the use of an isolator that can be made of flexible film, rigid metal framing, or solid plastic walls. Such isolators can be easily designed to house animal cages for larger laboratory animals such as ferrets, rabbits, or small NHPs. The cages kept within these isolators can be standard stainless-steel caging for rabbits or ferrets and also have squeeze back mechanisms as necessary. Multiple cages can be installed in a single isolator depending on the size of each cage. The isolator itself will have a HEPA-filtered intake and exhaust port, a magnehelic gauge, and a backup battery-powered motor available in the event of a power failure. In addition, the isolator can be fitted with 2 or more glove ports for the handling of animals. This represents an advantage in that the infected animal may stay within its home cage in the isolator and, when removed for procedures, it can still stay within the isolator in a procedural space in front of the cage and be accessible for manipulation using the portholes fitted with arm-sleeves and gloves; that is, the BSL-3–infected animal does not necessarily need to leave the isolator. Many different sized versions of these isolators have been in use for some time at the biodefense facilities at Porton Down, England.

These negative air pressure isolators can be of any size needed to house a range of cage sizes and animal species, and although they work well with somewhat larger laboratory animal species such as rabbits, ferrets, and small or young NHPs, they can be much more cumbersome and impractical for larger research animals.

With grant funding provided by the Department of Defense, the Mid-Atlantic Regional Center of Excellence for Biodefense and Emerging Infectious Diseases, and Project Bioshield, we prepared isolator units to work with the somewhat larger species noted above and to be able to infect those research animals with select agents. These agents included Bacillus anthracis, Francisella tularensis, Yersinia pestis, Coxiella burnetii, and Rickettsia prowazekii. To meet these new demands, adherence to several criteria was followed from the CDC publication Biosafety in Microbiological and Biomedical Laboratories 5^th Edition (BMBL).⁴ The BMBL states: “Biosafety Level 3 is applicable to clinical, diagnostic, teaching, research, or production facilities where work is performed with indigenous or exotic agents that may cause serious or potentially lethal disease through the inhalation route of exposure.”^4(p38) The BMBL further states: “The risk of infectious aerosols from infected animals or their bedding also can be reduced if animals are housed in containment caging systems, such as solid wall and bottom cages covered with filter bonnets, open cages placed in inward flow ventilated enclosures, HEPA-filter isolators and caging systems, or other equivalent primary containment systems. Actively ventilated caging systems must be designed to prevent the escape of microorganisms from the cage. Exhaust plenums for these systems should be sealed to prevent escape of microorganisms if the ventilation system becomes static, and the exhaust must be HEPA filtered. Safety mechanisms should be in place that prevent the cages and exhaust plenums from becoming positive to the surrounding area should the exhaust fan fail. The system should also be alarmed to indicate operational functions.”^4(p79)

We planned to meet all the above requirements by using designated ABSL-3 rooms that incorporated modified semi-rigid isolators to provide an additional level of containment of the agents and safety to the staff. The isolators were sized to be able to house caging for rabbits and smaller NHPs such as marmosets, African greens, and young macaques (eg, 1–3 kg). Additional criteria to be met included safety, maintenance, sanitation, illumination, security, and an emergency response plan. Approvals from the institution’s Environmental Health and Safety Program, Institutional Biosafety Committee, and the Institutional Animal Care and Use Committee were obtained. Some of the key factors in our development of these ABSL-3 projects with semi-rigid isolators included engineering controls for directional airflow and pressure differentials, redundant HVAC elements, HEPA-filtered exhaust, emergency power, multiple security layers, and support space.

COMPUTATIONAL FLUID DYNAMICS ANALYSIS OF ISOLATOR INTEGRITY

To verify the biocontainment and safety, assurance testing was required before actual studies with these agents could be initiated and performed. The impact of leaks, rips, and other semi-rigid isolator integrity issues needed to be evaluated through the use of numerical modeling and/or other predictive techniques before any formal ABSL-3 study was conducted. Approaches were undertaken to study potential breaches in containment as well as an assessment of the seriousness of those breaches. The breaches examined ranged from unlikely failures of the air-handling systems to accidental situations, including potential acts of deliberate sabotage.

METHODS

Computational Fluid Dynamics

The principles of computational fluid dynamics (CFD) were discovered well over 100 years ago, but practical numerical applications did not make their debut until the 1960s when digital computers became more commonplace. Today, computing power has increased so dramatically that fluid dynamics problems of the early days can now be solved instantaneously. In addition to speed improvements of the computing electronics, software developments have been made that allow large problems to be broken down into many smaller types of problems and solved in parallel.^5,6 CFD is a well-established and validated software tool that can verify, calculate, and visually display how very small particles such as viruses, bacteria, etc behave under certain conditions (eg, release within a containment device/system in a containment room) as a result of a BSL-3 breach of containment in the event of semi-rigid isolator rips, leaks, and other integrity losses. CFD has been used for a multitude of engineering applications such as the design of automobiles,⁷ ships,⁸ aircraft,⁹ building structures,^10,11 HVAC energy dynamics,^12–14 and biomedical engineering projects.^15,16 CFD has been successfully applied to the planning and design of laboratory animal facility projects, including air-handling systems, room airflow, canine kennel design, and ventilated caging applications and to solve/analyze problems that involve fluid flows.^17–23

CFD involves a software program analysis of fluid or gas flow. The concept of CFD is based on numerical software that subdivides an air space, such as a room, into many thousands or millions of small cubes (cells) of air. Each small volume is mathematically described by using the equations of the conservation of mass, momentum, and energy, which essentially says, “what goes in must come out.” If mass flows into 1 or more of the 6 faces of the small cube, for example, it must exit through 1 or more of the other remaining faces. The fundamental basis of CFD problems depends on the Navier–Stokes equation that accounts for the effects of air mass, momentum, energy, discrete phases (particles), pressure, and turbulence. After the room space is subdivided into its constituent cells, the CFD software solves for the behavior of each cell and its interaction with each surrounding cell, thereby simulating the overall behavior of the room.

Output from the CFD software is provided in several forms such as contours, vectors, or massless particle tracks. Contour outputs are read like a topographical map, but the lines indicate contours of constant pressure, temperature, or speed rather than altitude. Velocity vectors are color- and length-coded arrows, indicating both airflow speed and direction. Massless particle tracks visually define movement in 3-dimensional space.

Modified Semi-rigid Isolators

IsoTech Semi-Rigid Isolators (Envigo-Harlan) were used for containing the BSL-3 agents. Similar isolators are made by Class Biologically Clean, Ltd. and others. These units are most commonly used under positive pressure to assure that microbiological agents do not enter the isolator to contaminate the animals; this is known as a keep-out application of bio-exclusion. The isolators were modified as described below to maintain a negative pressure relative to the room and consequently contain the aerosolized BSL-3 select agents. This is a keep-in application of biocontainment. The terminology of keep-in keep-out was previously defined in laboratory animal science.²⁴

Three isolators were placed in 1 ABSL-3 room with a 100% filtered fresh-air inlet duct, centrally located in the ceiling, and 2 dedicated exhaust ducts at the rear of the room, near the ceiling, that passed through pre-filters and HEPA filters prior to exhausting through ducts to outside air. Personnel in scrub suits entered through an airlock that is connected to the holding room suite. Visual airflow monitors allowed the staff to ascertain when the anteroom was negative to the entrance hallway. In addition, there was a room pressurization panel located at the entrance of the facility that visually and audibly informed staff of the current room/space pressures prior to entry. A green light visual indicator relayed negative pressure, whereas a positive pressure was relayed by a red light and audible alarm.

While in the anteroom, each employee donned personal protective equipment (PPE) of breathable full-body Tyvek coveralls, shoe covers, a hair bonnet, and a powered air purifying respirator. When passing through the anteroom, visual airflow monitors allow the staff to know that the animal holding room is negative to the anteroom. The containment housing/semi-rigid isolators provided a first barrier, followed by a secondary barrier of PPE, and the ABSL-3 room provided a tertiary barrier.

An advantage of the IsoTech Semi-Rigid Isolator is the relatively small footprint (20.6 square feet). The isolators were 38 inches wide, 78 inches long, and 76 inches high. The overall size of these isolators allows 3 units per ABSL-3 room with 156 square feet. This can permit up to 3 different ABSL-special agent studies in 1 room. Isolators can be custom made to any size required.

The IsoTech isolators were fabricated with a durable, transparent polyvinyl chloride (PVC) membrane (20 mil) allowing for unobstructed sight as well as being resistant to tears, punctures, and rips.²⁵ The PVC was secured on a frame of 1.5-inch anodized and extruded aluminum. The isolators were modified for negative air flow (pressure) that included multiple back-up systems to assure their integrity and function for studies with BSL-3 select agents. Intake air was filtered through a pre-filter and then a HEPA filter, whereas the exhaust air was filtered through 2 in-line HEPA filters. The exhaust fans were a double redundant system; if 1 motor failed, then a second motor immediately started using a lead/lag system. Negative pressure is not lost. In case of a motor failure, there are audible and visual alarms, and a short message service text message is sent to the cell phone of the containment animal facility manager and veterinary resources key personnel. The motors were connected to the research animal facility electrical system that has a back-up generator in case of electrical outage. Each isolator had a dedicated battery that is triggered into operation in the event of complete facility electrical outage. With these redundant electrical systems, there was assurance of continuous energy supply to the exhaust motors. The uninterrupted negative airflow is essential to the containment of the aerosolized BSL-3 select agents. Each fan motor was wired in such a way as to prevent positive pressurization; so, the fans are either running negative when powered or turned off if all levels of protection fail. In the event of dual fan failure, the isolator would revert to static mode. However, because there are both supply and exhaust HEPA filters to contain the agent within the enclosure, any aerosolized agent would be blocked from escape.

Each room was equipped with a Class III Biological Safety Cabinet. Movement of animals between the semi-rigid isolator and Class III BSCs is accomplished using transportable semi-rigid acrylic glass (eg, Plexiglas) drums. Sterilization of the entry port is accomplished using chlorine dioxide (20.8% Sodium Chlorite: MB-10 Quip), a proven sterilizing agent that kills bacteria, fungi, algae, and viruses.²⁶ Each end of the drum is fitted with a flexible film cover and taped in place. Exposure of the animals to the aerosolized agents is accomplished inside the Class III BSCs. The animals are placed inside a conical, full-body Plexiglass holder. The apex of the body holder fits securely into a nebulizer. A low flow of 2–6 L/min is effective for delivering a dose of several million organisms. This shape allowed comfortable restraint assuring nose-only exposure and prevented any of the aerosolized agent from spilling over to contaminate the animal’s body surface. Within 2 weeks, the animals were adapted to this training regimen using a placebo aerosol. The few animals that were unable to adapt to the prescribed training procedure were removed from the study.

To evaluate the integrity of the semi-rigid isolators, the most likely failure mechanism we considered was the loss of integrity of the flexible film itself. As stated earlier, an exhaust motor malfunction was rare because of the in-line redundant exhaust motor. An emergency electrical generator was available in the building as well as each isolator having a dedicated and fully charged battery backup.

Case Studies

Different case study failures were simulated using CFD. Categories included:

1. Small rips/separations of seams;

2. Large rips of 6 inches;

3. Hatch opening due to the transport drum falling off or not being replaced properly. There is an inside seal on each hatch opening as well as an outside seal. These seals are kept in place with metal clamps. The outside seal is unsnapped when the transport drum is connected and then re-clamped. Even if the transport drum fell off, the inside seal would remain intact because it is clamped. For this CFD demonstration, both the inside seal with its clamp and outside seal with its clamp were simulated as removed. This is a hypothetical situation to evaluate the CFD consequences of both inside and outside seal removal.

Each of these 3 categories was evaluated with:

• The exhaust motor properly working: ie, negative air flow

• The exhaust motor not working: ie, no air flow

To assess each scenario, a simulated mass flow of nebulized droplets introduced into the interior of the isolators was monitored for both their migration, airflow patterns, and ultimate demise of the droplets. The paths taken by each of the droplets, or groups of droplets, as they migrated through the airspace were visualized using “mass particle tracks” generated by the CFD software. This visualization technique allows the user to see the path taken by the droplets from source to a stopping point in the form of 3-dimensional lines passing through the airspace.

RESULTS

1. With the exhaust fan operating normally (negative pressure) and a small breach in the isolator membrane, such as a rip or seam separation, no droplets were observed entering the ABSL-3 room (Figure 1). All mass particle tracks were completely contained inside the isolators. No particles escaped into the room.

2. With the exhaust fan stopped (neutral pressure) and a small breach in the isolator membrane, such as a rip or seam separation, only a few droplets were observed entering the ABSL-3 room (Figure 2). The vast majority of mass particle tracks were completely contained inside the isolators. The tracks that did escape were quickly scavenged by the room exhausts.

3. With an exaggerated rip of 6 inches (15 cm) in the isolator membrane, representing sabotage by someone willfully cutting the PVC with a knife, and the exhaust fan operating normally (negative pressure), no droplets were observed entering the ABSL-3 room (Figure 3). All mass particle tracks were completely contained inside the isolator. No particles escaped into the room.

4. With the normal operation of the isolator exhaust system plus the loss of the large outer hatch seal, and the rare event of the inside seal removed (Figure 4), no droplets were observed entering the ABSL-3 room. All of the mass particle tracks were adequately contained within the isolators. This is a hypothetical situation because the inner seal is secured with a metal clamp. The drum would only fall off if accidentally dropped. Once the drum is in place, even if inadvertently not clamped, it would stay in place.

5. Another hypothetical situation involved the failure of the isolator exhaust system along with the loss of the outer hatch plus the rare event of inside seal removal. (Figure 5). The drum could fall off if accidentally dropped. This would lead to a significant breach of containment. According to CFD analysis, the room exhaust was unable to keep up with the leakage rate from the isolators, and this resulted in suspended droplets in the airspace of the room. Yet even if this occurred, the fail-safe engineer blower systems and alarms would quickly return the blowers to a negative airflow, and the room would be promptly decontaminated by staff wearing PPE.

6. The motor was turned on to positive airflow with a small rip (Figure 6). The purpose of the CFD example is to visualize mass particle loss from the semi-rigid isolator into the room. NOTE: This could NOT occur due to the electrical wiring of the motor that prevents the motor from producing positive airflow.

Figure 1 .

Schematic (diagonal view) of the modified semi-rigid plastic isolator with a small rip or plastic seam opening with negative pressure. The lines represent the migration and airflow patterns or mass particle tracks of nebulized droplets. All mass particle tracks were completely contained inside the isolators. No particles escaped into the room.

Figure 2 .

Schematic (overhead view) of the modified semi-rigid plastic isolator with a small rip or plastic seam opening with no exhaust flow/neutral pressure. The lines represent the migration and airflow patterns or mass particle tracks of nebulized droplets. The vast majority of mass particle tracks were completely contained inside the isolators. The tracks that did escape were quickly scavenged by the room exhaust.

Figure 3 .

Schematic (diagonal view) of the modified semi-rigid plastic isolator with a large rip and with negative pressure exhaust. The lines represent the migration and air flow patterns or mass particle tracks of nebulized droplets. All mass particle tracks were completely contained inside the isolator. No particles escaped into the room.

Figure 4 .

Schematic (diagonal view) of the modified semi-rigid plastic isolator with a port opening due to the transport drum falling off by not being properly secured with the clamps, and with negative air flow. The lines represent the migration and air flow patterns or mass particle tracks of nebulized droplets. All of the mass particle tracks were adequately contained within the isolators.

Figure 5 .

Schematic (overhead view, only 1 of 3 isolators affected) of the modified semi-rigid plastic isolator with port opening due to the transport drum falling off or not being properly replaced and clamped, and with no exhaust flow/neutral pressure. The lines represent the migration and air flow patterns or mass particle tracks of nebulized droplets. The room exhaust was unable to keep up with the leakage rate from the isolators, and this resulted in suspended droplets in the air space of the room. The failsafe engineer blower systems and alarms would quickly return the blowers to a negative airflow, and the room would be promptly decontaminated by staff.

Figure 6 .

Schematic (diagonal view) of the semi-rigid plastic isolator with a small rip and with positive airflow. The lines represent the migration and air flow patterns or mass particle tracks of nebulized droplets. The purpose of this CFD example is to visualize mass particle loss from the semi-rigid isolator into the room. NOTE: This could not occur due to the electrical wiring of the motor that prevents the motor producing positive airflow.

CONCLUSION

The CFD testing demonstrated that with the exhaust motor running properly with a negative air flow, the containment system worked effectively and safely even with different losses of integrity features of the semi-rigid isolators. This was proven with a small rip, large rip, or transport drum cover falling off or drum cover not being replaced. The results provided assurance of the semi-rigid isolator’s biocontainment features. The only way that the biocontainment failed was with 2 simultaneous failures: malfunction of the exhaust motor and a loss of integrity of the semi-rigid isolator. In practicality, the motor malfunction would be avoided due to a second exhaust motor being able to function in case the first exhaust motor fails. Alerts and alarms occur with the loss of the first exhaust motor. A replacement motor can be expeditiously installed. Meanwhile, the second motor would function properly on its own. Loss of the exhaust motor due to an electrical failure in the facility immediately triggers the diesel-powered back-up generator that supplies electricity to the building to come on. As a further back-up, each isolator has a dedicated and fully charged battery. Even with the improbable event of these 2 failures, the ABSL-3 room provides exhaust of the infectious agents and personnel were fully protected using PPE. This arrangement provides biosafety to the staff, the facility, and—with the HEPA filters of the room exhaust—safety to other rooms within the facility as well as prevention of the BSL-3 agent release into the external environment. Biocontainment was effective unless 2 failures occurred simultaneously: malfunction of the exhaust motor and loss of integrity of the flexible film isolator. The CFD results substantiated the use of the semi-rigid isolators as described with the facility and back-up engineering systems for use with BSL-3 agents.