oxygen supplies during a mass casualty situationoxygen supplies when the demand is suddenly...

Oxygen Supplies During a Mass Casualty Situation

Ray H Ritz RRT FAARC and Joseph E Previtera RRT

IntroductionCompressed Gas CylindersOxygen ConcentratorsLiquid Oxygen DevicesBulk Liquid Oxygen SystemsDeveloping an Oxygen Resource Management PlanOxygen Conserving OptionsSummary

Mass casualty and pandemic events pose a substantial challenge to the resources available in ourcurrent health care system. The ability to provide adequate oxygen therapy is one of the systemsthat could be out-stripped in certain conditions. Natural disasters can disrupt manufacturing ordelivery, and pandemic events can increase consumption beyond the available supply. Patients mayrequire manual resuscitation, basic oxygen therapy, or positive-pressure ventilation during thesescenarios. Available sources of oxygen include bulk liquid oxygen systems, compressed gas cylin-ders, portable liquid oxygen (LOX) systems, and oxygen concentrators. The last two are availablein a variety of configurations, which include personal and home systems that are suitable forindividual patients, and larger systems that can provide oxygen to multiple patients or entireinstitutions. Bulk oxygen systems are robust and are probably sustainable during periods of highconsumption, but are at risk if manufacturing or delivery is disrupted. Compressed gas cylindersoffer support during temporary periods of need but are not a solution for extended periods oftherapy. Personal oxygen concentrators and LOX systems are limited in their application duringmass casualty scenarios. Large-capacity oxygen concentrators and LOX systems may effectivelyprovide support to alternative care sites or larger institutions. They may also be appropriateselections for governmental emergency-response scenarios. Careful consideration of the strengthsand limitations of each of these options can reduce the impact of a mass casualty event. Key words:oxygen, compressed gas, oxygen concentrators, liquid oxygen, bulk oxygen, mass casualty, emergencypreparedness, pandemics. [Respir Care 2008;53(2):215–224. © 2008 Daedalus Enterprises]

Introduction

Among the many challenges that mass casualty eventscreate is the reliable delivery and maintenance of adequate

oxygen supplies when the demand is suddenly increased.Cylinder and bulk oxygen stores can become strained, anddelivery and restocking may be interrupted. During theseevents, large numbers of patients may need a wide varietyof interventions, including manual resuscitation, transportsto and within care sites, basic oxygen therapy, and me-chanical ventilation.

Mass casualty scenarios can be either regional or global.Regional events such as earthquakes, floods, explosions,

Ray H Ritz RRT FAARC and Joseph E Previtera RRT are affiliated withthe Department of Respiratory Care, Beth Israel Deaconess Medical Cen-ter, Boston, Massachusetts.

Mr Ritz presented a version of this paper at the 40th RESPIRATORY CARE

Journal Conference, “Mechanical Ventilation in Mass Casualty Scenar-ios,” held July 16-17, 2007, in Reno, Nevada.

The authors report no conflicts of interest related to the content of thispaper.

Correspondence: Ray H Ritz RRT FAARC, Department of RespiratoryCare, Beth Israel Deaconess Medical Center, 330 Brookline Avenue,Boston MA 02215. E-mail: [email protected].

RESPIRATORY CARE • FEBRUARY 2008 VOL 53 NO 2 215

and fires affect a city or region and can disable or over-whelm operations in hospitals and medical response sys-tems. Though hospitals and regions can be supported fromoutside the affected area in time, disruptions that disablehospital infrastructures, such as bulk oxygen systems orthe piping system that connects to the patient-care areas,can have long-term effects. A pandemic event is morefrightening since its effect would be global and resourcedemand could far outstrip any preparations or responseplans. Workforce reductions could reduce manufacturingcapacity and interfere with delivery. Large numbers ofpatients who require oxygen could quickly deplete normaloxygen supplies and render life-support equipment use-less.

In both local catastrophes and pandemic events, hospi-tals may be overwhelmed, which would force an expan-sion of patient care to areas not designed for that purpose,such as auditoriums, dormitories, schools, and other avail-able sites. Government agencies may deploy mobile fieldhospitals that are equipped with a variety of critical sup-plies, but not all may include utilities such as bulk oxygen.Careful planning at state and local levels as well as withinindividual hospitals could help support the oxygen require-ments of alternative care sites.

Oxygen is available in compressed gas cylinders, por-table liquid oxygen (LOX) systems, produced by oxygenconcentrators, and supplied by bulk LOX systems. Thepurpose of this paper is to narrowly focus on options tobest utilize resources and sustain the ability to provideoxygen to patients during mass casualty events.

Compressed Gas Cylinders

Compressed oxygen cylinders come in a variety of sizes(Table 1) and can be classified into 2 general types: smallerportable cylinders (sizes B through M), and larger cylin-ders (sizes H, K, and larger), which are best suited forstationary placement. The gas capacities of the most com-monly available cylinder sizes are also listed in Table 1.With any given cylinder size, the gas capacity may differslightly, depending on the cylinder manufacturer.

Portable cylinders are commonly used to provide oxy-gen during ground and air transports, for moving patientswithin hospitals, and to deliver oxygen in the field. Theyare convenient because of their light weight and small size,but are limited in the amount of oxygen they contain.Because of this limitation, portable cylinders do not offera long-term solution for supplying oxygen to the criticallyill patient. Smaller portable cylinders (sizes D and E) arecommonly used for patient transports, since they are light-weight and compact and under most conditions will supplyoxygen for 30 min to one or more hours, depending on theflow rate of gas delivery.

Transport cylinders today are made of steel, aluminum,or composite materials. Steel cylinders are the most eco-nomical but are heavier and pose a substantial risk inhospitals of being inadvertently exposed to magnetic res-onance imaging areas, where they can become lethal pro-jectiles. They are generally being replaced with their alu-minum counterparts. Aluminum cylinders are moderatelypriced, light-weight, and can be manufactured to be safe ina magnetic resonance imaging environment. Although com-posite cylinders are extremely light-weight, their expensegenerally makes it cost prohibitive to be commonly usedin a medical environment.

Larger cylinders such as size H or K are most com-monly constructed of steel (but are available in aluminum)and provide a greater reservoir of gas. They are commonlyused as a stand-alone oxygen source in areas where walloxygen is not available. They are also used in groups,where a series of large cylinders are connected together bya manifold to either support specific areas or serve as anemergency reserve to a main supply. One possibility whenmass casualties cause a sudden increase in patients is toconvert areas such as cafeterias or auditoriums into pa-tient-care areas by deploying a pre-constructed manifoldto link large cylinders to supply oxygen. Either commer-cially available or easily fabricated on-site, a system oflarge cylinders, gas regulators, high-pressure hose, andflow meters could be quickly deployed to supply oxygento multiple patient sites. An example of a commerciallyavailable system is depicted in Figure 1 and Figure 2.These systems can be assembled in a variety of configu-rations; however, one must consider the flow limitation ofthe regulator on the manifold. If a regulator is only capableof providing a total output of 150 L/min, the number offlow meters multiplied by their flow rate setting is limitedto a combined output of 150 L/min.

Traditionally, cylinders are filled to a service pressurebetween 2,000 and 2,200 pounds per square inch (psi).Recently introduced higher-pressure cylinders can be pres-

Table 1. Compressed Gas Cylinder Specifications*

CylinderType

SizeService

Pressure (psi)CubicFeet

Liters

Portable B 2,015 5.8 164C 2,015 8.8 249D 2,015 14.7 416E 2,015 24 679E (HP) 3,000 35 991M 2,015 110 3,113

Fixed H 2,015 220 6,226K 2,015 266 7,528

*Commonly found portable and stationary oxygen cylinders. Note that a cylinder’scompressed gas volume may differ slightly depending on the manufacturer of the cylinder.

OXYGEN SUPPLIES DURING A MASS CASUALTY SITUATION

216 RESPIRATORY CARE • FEBRUARY 2008 VOL 53 NO 2

surized up to 3,000 psi, which increases the availableamount of oxygen in a size E cylinder from approximately650 L to approximately 1,000 L. They require a regulatorrated for this higher pressure but otherwise operate thesame as the conventional 2,000-psi cylinders.

One option available for portable cylinders is to havethem supplied by the vendor with a regulator attached toeach cylinder. Though this increases the cost of the cyl-inder service contract, it has the advantage during masscasualty situations of ensuring a more adequate supply ofregulators. Using oxygen cylinders with an integrated valve,regulator, and flow meter also eliminates the possibility ofa fire or explosion by civilian medical volunteers who maynot be trained in the correct procedure for changing oxy-gen cylinder regulators. In a joint advisory issued April2006, the National Institute for Occupational Safety andHealth and the U.S. Food and Drug Administration1 cau-tioned that “plastic crush gaskets never be reused, as theymay require additional torque to obtain the necessary sealwith each subsequent use. This can deform the gasket,increasing the likelihood that oxygen will leak around theseal and ignite.” The Department of Homeland Security,as a result of September 11th 2001, Hurricane Katrina,tsunamis, possible flu pandemic, and other incidents, hasprovided funds for communities to establish citizen Med-ical Reserve Corps. These volunteer groups may have var-ious degrees of familiarity with the proper handling ofoxygen cylinders or specifically changing the oxygen cyl-inder regulators. This potential for mishap is eliminated byusing oxygen cylinders with integrated valves, regulators,and flow meters.

The higher service fee of having cylinders supplied withregulators will be partially offset by a reduction in both thenumber of replacement regulators purchased each year anda reduction in repair costs. If this option is taken, oneshould consider stockpiling some of the hospital-ownedregulators being taken out of service, because in the caseof a mass-casualty scenario you could receive additionalportable cylinders from vendor stockpiles that require reg-ulators.

There are several “choke points” associated with relyingon compressed gas cylinders to supply oxygen to patientsduring mass casualty scenarios. Oxygen vendors have afixed supply of cylinders and rely of hospitals and otherusers to promptly return empty stock for refilling. Failureto gather and return cylinders will lead to a shortfall insupplies. Refilling cylinders, though efficient, does requirea fair amount of manual activity. During a pandemic eventthis activity could be compromised by a reduction in ven-dor workforce.

Vendors have stockpiled older but still serviceable cyl-inders for emergency use. For example, in 2005, in re-sponse to the 2004 Hurricane Ivan, LifeGas implementeda Hurricane Inventory Preparedness Plan, which allowedcertain hospitals to reserve a limited number of additionalE cylinders during the 6-month southeastern United Stateshurricane season (personal communication, DebbieCapuano, Pandemic Flu Coordinator, BOC Gases, MurrayHill, New Jersey, 2007). Out of necessity, the Hurricane

Fig. 1. Portable oxygen delivery manifold 1. A portable manifold todeliver oxygen at an alternative care site includes a master regu-lator (bottom) that connects to an oxygen cylinder and a series offlow-regulation devices connected in sequence via high-pressuretubing. (Courtesy of Flotec, Indianapolis, Indiana.)

Fig. 2. Portable oxygen delivery manifold 2. An alternative portableoxygen delivery manifold utilizes nipple adapters that are actuallyrestrictors that control the flow rate to the patient. They are avail-able in color-coded versions that make it easy to distinguish var-ious flow rates. (Courtesy of Flotec, Indianapolis, Indiana.)



Inventory Preparedness Plan stipulated the timely return ofused cylinders to sustain the system. Other vendors havedeveloped similar stockpiling plans to utilize older or sur-plus cylinders during emergencies. In a pandemic scenariothere is also the concern of disinfecting used cylindersprior to returning to the vendor for refill. Who would dothat: vendor or hospital staff? Though this seems like asimple task, given the need to contain the spread of con-tagion and the anticipated reduction in workforce, theremay be inadequate staffing to collect and disinfect thesecylinders.

Another logistical issue associated with cylinders is theavailable storage space in hospitals. Commercially avail-able emergency-response cylinder-storage systems areavailable, and an example is depicted in Figure 3. Thissystem maintains approximately 80 conventional E cylin-ders and a supply of regulators in a wheeled cabinet, andcould support a limited emergency event. Since thissystem contains over 1,800 cubic feet of compressedgas, certain storage regulations apply. Cylinder storagemust conform to certain National Fire Protection Associationsafety standards2 and possibly other state and local guide-lines. These standards may include fireproofing requirements,ventilation standards, and cylinder handling requirements,based on the volume of gas stored:

• � 3,000 cubic feet requires lockable doors, storage racksor chains, indirect heat (if required), and dedicated ven-tilation systems

• � 300 cubic feet to � 3,000 cubic feet requires lockabledoors and passive ventilation

• � 300 cubic feet (12–13 standard E cylinders) may bestored without any restrictions as an operational supplyto support clinical operations

Because of space limitations within most institutionsand the regulatory requirements associated with cylinderstorage that impart an additional expense to hospitals, ad-ditional space is often not available for substantial quan-tities of disaster supplies. With these storage issues inmind, careful planning must be done and, in reality, areliance on compressed gas cylinders is not a long-termsolution for pandemic events.

Oxygen Concentrators

Oxygen concentrators range in size from small portablepersonal units to large industrial systems that can producetons of oxygen per day. They all utilize a sieving processto extract oxygen from air and provide � 90% oxygen ata variety of flow outputs. Compressed air is applied to azeolite granule bed that absorbs nitrogen, leaving a resid-ual high concentration of oxygen. By switching to a sec-ond, separate bed of zeolite, the extracted oxygen can bediverted to a holding chamber and the nitrogen is dissi-pated from the saturated first zeolite sieve. The process ofswitching between 2 sieves is termed pressure swing ab-sorption and is continuously repeated to provide a contin-uous flow of nearly pure oxygen. Depending on the sizeand model, oxygen concentrators can deliver oxygen of� 90% at various flow rates. Generally, the greater thepurity of the oxygen generated, the less liters per hour canbe generated.

The traditional medical application of oxygen concen-trators has been to support low-flow oxygen therapy forhome-care patients. Cabinet-size systems and smaller por-table devices have provided flows up to 15 L/min (de-pending on the model) and serve home application well.They are not commonly available in most hospitals be-cause under normal conditions the institutional bulk liquidsystem is more economical and convenient. The questionis, do they have a role in mass casualty events? Personaloxygen concentrators have an advantage over cylinder gasbecause they generate their own supply of oxygen and donot need to be refilled. This avoids the logistical issues ofmonitoring cylinder levels, changing depleted cylinders,and exchanging empty cylinders for full ones. They dorequire a source of electrical power and are less conve-nient than an oxygen flow meter connected to a wall out-let. They are expensive and large relative to oxygen flowmeters, and stockpiling them for use in a mass casualtyevent is probably not cost-effective for most hospitals.

Fig. 3. Oxygen cylinder stockpiling system. Oxygen cylinders canbe efficiently stockpiled in a convenient mobile storage systemthat also contains regulators and basic patient interfaces. It shouldbe noted that once the total volume of gas exceeds 300 cubic feet,additional storage requirements must be followed. (Courtesy ofAirgas Puritan Medical, Radnor, Pennsylvania.)



There are larger oxygen-generating systems availablethat range in a variety of sizes from portable trailer-baseddesigns to fixed systems that can provide the oxygen needsof an entire hospital.3 A search of the Internet yielded thefollowing list of oxygen-generator manufacturers suitedfor hospital applications:

• AirSep, Buffalo, New York• Generon Innovative Gas Systems, Houston Texas• Oxygen Generating Systems Intl, North Tonawanda, New

York• Oxair Ltd, Niagara Falls, New York

Depending on the size, these oxygen generators canproduce substantial amounts of nearly pure oxygen. Forexample, the OG-5000 (Oxygen Generating System Intl,North Tonawanda, New York) can produce 5,000 standardcubic feet per hour (over 1,000 liquid gallons/d) of 93%oxygen and can support wall oxygen delivery systems thatprovide the standard 50 psi working pressure. In reality, itis likely that the exact size and production capacity can bedesigned to meet any specific application. Oxygen gener-ators of this size include air compressors and gas dryers toprovide a flow of compressed air to a set of extractionsieves that feed a reservoir tank that supplies the patient-care site. They require moderate to large amounts of elec-tricity that can be provided by either a public utility or aportable generator. In addition to being able to supplyadequate oxygen flow for wall oxygen systems, they canalso refill cylinders. These systems could offer an attrac-tive option for hospitals that may be at risk of oxygen-delivery interruptions that could occur during hurricanesor other natural disasters—as long as electricity is avail-able to power them. Trailer-based systems may be an at-tractive option for federal, state, and local governments tosupport alternative care sites that could be deployed duringmass-casualty events. They are mobile, and, with an ap-propriate set of portable manifolds, numerous patients couldbe quickly supported.

Large oxygen-generating systems do have several draw-backs. They are physically large and require as much ormore physical space as a bulk LOX system. They must beprotected from environmental extremes. This means theymust be installed indoors or fitted with some type of pro-tective structure. They are not “off the shelf” systems, andeach one must be designed and built according to individ-ual specifications. The cost for a system that could supportan entire hospital would probably match or exceed the costof a bulk liquid system. The cost for the required electric-ity could be substantial. And since each system is designedand built for a specific application, the time from orderinga system until its installation could be substantial.

Liquid Oxygen Devices

LOX, both portable and satellite filling stations, havebeen frequently utilized in both home and hospital set-tings. These devices are essentially large insulated bottles,and for their weight and size, they offer substantially greateroxygen capacity than compressed gas cylinders. Home-usemodels generally provide only low-flow oxygen (0.25–15 L/min), but larger systems can support 50-psi gas re-quirements. Single-patient LOX systems do not offer adistinct advantage unless they are already in wide use in ahospital. Most hospitals do not employ portable, single-patient LOX systems in place of compressed gas cylinders,for several reasons. LOX systems, when not in use, slowlydissipate their contents to the atmosphere as the liquid gasevaporates, in order to maintain a stable pressure inside thechamber. Second, portable LOX systems require periodicrefilling from a satellite reservoir station, and although notdifficult, this process requires a modest amount of train-ing.

There are large-capacity mobile LOX systems that areavailable in several different configurations (Fig. 4) tosupport alternative care sites and refill home systems. Thistype of trailer or truck-based system can provide relativelylarge amounts of LOX to prefabricated portable distribu-tion systems. These are simply mobile versions of bulkoxygen systems that supply multiple patient-distributionsystems. An example of this type of oxygen delivery sys-tem is shown in Figure 5 (available from Essex Cryogen-ics, St Louis, Missouri). This system utilizes a 75-liquid-liter mobile LOX container (� 60,000 gaseous liters ofoxygen) that is connected to a series of patient-deliverykits. Each patient-delivery kit is configured with a set of10 high-pressure hoses and can deliver oxygen to 10 pa-tients, at flow rates of up to 15 L/min. The patient-deliverykits can be connected in sequence to extend the systemuntil the maximum flow delivery of the LOX system isreached. Oxygen flow to each patient can be regulated bya separate flow-control valve located in the patient-deliv-

Fig. 4. Liquid oxygen (LOX) refilling trailer. Field LOX systems canbe supported by such conveniently sized, trailer-mounted LOXreservoirs. (Courtesy of Essex Cryogenics, St Louis, Missouri.)



ery kit. This type of system provides federal, state, andlocal agencies, as well as hospitals, with an option for theirdisaster plans.

The limitation of all LOX systems, besides the evapo-rative loss when not in use, is the logistical aspect ofresupply. Disasters such as floods and earthquakes caninterrupt transportation and delivery of LOX. Pandemicevents could substantially reduce the workforce at gas-separation plants and impact production and delivery. Op-erating a mobile LOX system requires regular training ofan adequate number of individuals to ensure its function-ality.

Bulk Liquid Oxygen Systems

Most hospitals utilize bulk LOX systems to providewall oxygen at approximately 50 psi to patient-care areas.These systems are robust and reliable under most condi-tions. The container contents are usually monitored eitherby the vendor electronically or by the hospital mainte-nance department (or both). They are often refilled in offhours by large liquid-gas transport trailers to avoid con-gesting the area around the facility. The vendor deliverystaff are trained and certified in the filling technique. Assuch, respiratory therapists tend to take these systems forgranted, much the same as we take most utilities for granted.In planning for mass casualty and pandemic events it is

important to understand the operation and potential limi-tations of the institution’s bulk oxygen system

Like portable LOX systems, bulk oxygen systems arecomposed of an inner vessel and outer vessel. The spacebetween is maintained as a continuously monitored vac-uum to insulate the LOX, which helps maintain vesselpressure and reduces the evaporative loss. The inner vesselis maintained at a pressure near 130 psi. LOX is with-drawn from the vessel and routed through a vaporizer thatconverts it to a gas, and then through a series of regulatorsthat reduce the pressure down to the hospital’s workingpressure of approximately 50 psi. The vaporizer is thefinned structure in the bulk system that can be covered inice. There are usually 2 vaporizers, and the system auto-matically switches between the two as the ice reduces itsefficiency. An economizer system collects the evaporatedgas from the main vessel as the pressure rises and routesthis recovered gas through the vaporizer and into the flowinto the hospital. Under normal conditions, a hospital’soxygen consumption is sufficient to prevent any gas lossinto the atmosphere, as may occur with personal LOXsystems. There are also several pressure-relief valves thatprevent excess pressure from developing in the system.

Each institution’s bulk oxygen system is configured toits specific needs, and it is beyond the scope of this paperto thoroughly detail the process for filling the main sys-tem. There is a rigorous and thorough training process that

Fig. 5. Mobile oxygen distribution system (MODS) and patient distribution kits in “string-along” setup. Mobile liquid oxygen (LOX) systemscan provide oxygen for multiple patients. This model contains 75 L of LOX (64,000 gaseous L) and can be connected to a series ofpatient-distribution kits, each of which can deliver oxygen to 10 patients. (Courtesy of Essex Cryogenics, St Louis, Missouri.)



must be completed to achieve certification to fill and main-tain them. This includes appropriate hazardous-materialtraining as well as practical training on filling the vesselwithout allowing excess pressure to develop. That beingsaid, there is a concern that during a pandemic event suchas an outbreak of avian flu (with an attack rate of 35%) thevendor workforce could be disrupted, and trained person-nel may not be available to resupply the bulk system.Vendors have substantially more cylinder-filling plants thanthey have gas-separation plants. Gas-separation plants dif-fer in size, but the number of delivery technicians anddelivery tankers is limited. It may be appropriate for in-stitutions to certify some of their engineering staff in therefilling process in case of such an event.

The diagram displayed in Figure 6 is a basic version ofa bulk oxygen vessel. The following is a substantiallysimplified version of the refilling process, adapted fromthe Taylor-Wharton operation manual for vertical storagetanks.4 After confirming that the trailer contents matchthose of the receiving vessel’s, the supply tanker is con-nected to the bulk system’s main inlet valve. The bulksystem has a bottom fill valve and a top fill valve. Thebottom fill valve is opened, and then the top fill valve. Asflow through the bottom fill valve increases the liquidvolume in the vessel, it compresses the evaporated gas inthe top of the vessel. By filling simultaneously from thetop of the vessel, using the top fill valve, gaseous oxygenis reliquefied and the vessel pressure is reduced. Thus, byadjusting the bottom fill valve and the top fill valve, the

vessel pressure can be maintained as it is filled. Once thevessel is three-quarters full, an overflow valve (the trycockvalve) is opened. When LOX spurts from the trycock valve,the vessel is full and the main inlet valve is closed. Allvalves are now closed, and the filling lines are allowed todepressurize before being disconnected. There are otheressential steps not described in this paper, and it is imper-ative that untrained, uncertified personnel should not at-tempt this procedure.

Gas-separation plants currently operate at nearly full ca-pacity. There are periodic reductions in their production, dueto seasonal industrial consumption or to limit power con-sumption during peak electrical use. Most production plantsstore several days or more of product in on-site vessels, sothese variances in production do not affect delivery to con-sumers, but there is a limit on the volume of additional LOXthat can be produced at current facilities. Diversion of LOXfrom industrial applications can be done during extreme con-ditions. One potential method to increase LOX delivery tohospitals is to convert nitrogen tankers to oxygen. This iseasily done by converting the connection fixtures on the tankerto match the gas (if the conversion hardware is readily avail-able), and flushing the nitrogen tanks with oxygen beforefilling. Under all conditions, special care must be used toconfirm that the correct gas is being loaded into the bulksystem. Though delivery tankers are equipped with oxygenanalyzers to verify that the contents are correct, there havebeen a number of documented contaminations of bulk sys-tems with incorrect gases.5

Fig. 6. Bulk oxygen storage vessel. A typical bulk oxygen storage vessel is designed to fill from both the bottom and top, in order to maintaina stable vessel pressure during the filling process. P � vessel pressure monitor. V � vessel liquid oxygen volume monitor.



Bulk oxygen systems provide adequate reserves under mostconditions. Unless disabled by accidents or natural catastro-phes, they are probably able to provide the needed supplyduring short-term events. Their primary risk is during long-term high-consumption periods. Since deployment of oxygendelivery systems in alternative care sites presents a markedlogistical challenge, most hospitals will plan to move non-oxygen-dependent patients to these locations (if needed) first,reserving bed space with robust oxygen availability for thosepatients who require supplemental oxygen.

Developing an Oxygen Resource Management Plan

Developing an oxygen resource management plan willdepend on the response contingencies an institution plansto implement. Our institution has planned to develop analternative care site in a nearby university dormitory. Weplan to reposition in-patients with less acute illness and nooxygen needs to this location in order to free on-site bedsfor those patients who need oxygen and other more tech-nically demanding interventions. This would mean thatour proposed alternative care site would have modest ox-ygen resource demands, and we plan to support it with alimited number of emergency oxygen cylinders.

We also plan on utilizing public waiting areas as triagesites to support overflow from our emergency department.These areas would be supported with portable cylinderoxygen, with the intent to move patients to a more stableand supportable location as soon as possible. The patientcapacity of a surge area and its anticipated duration of usewill dictate the number of additional cylinders that shouldbe stockpiled. We allocated 3 E cylinders per bed space tosupport this triage area.

It is our philosophy that long-term (more than severalhours) oxygen therapy from cylinders is not a feasibleoption, and patients should be moved to a location wherea more robust oxygen supply is present. In these morepermanent patient-care areas we attempt to maintain one

oxygen flow meter at each potential bed site. Oxygen flowmeters are not prohibitively expensive, and maintainingthem at potential bed sites avoids the need to distributethem during an emergency. Additional units are storedon-site to replace lost or broken units. Although moreexpensive than standard flow meters, flow meters withadditional high-pressure outlets increase the number andflexibility of wall oxygen outlets. Regulators for both por-table and nonportable oxygen cylinders should also bestored in reserve, and the number should be based on thepotential surge capacity.

We also assessed the ability of our bulk LOX system tomeet the demands potentially placed on it by a mass ca-sualty event. We did a spot survey of oxygen utilization onone campus at Beth Israel Deaconess Medical Center. OurWest Campus facility consumes an average of 450 gallonsof LOX per day (1.5 million gaseous liters per day) and issupported by a 9,000-gallon bulk vessel and a 1,500-gal-lon reserve vessel. This system is resupplied every 5–10days, depending on the vendor’s schedule. Our West Cam-pus has 385 beds, with the distribution displayed in Ta-ble 2. By gathering data on a random sample of patients,we estimated the amount of oxygen being used in variousareas. All ICU patients were assessed for their oxygen use,and the status of their manual resuscitator was noted todetermine if it had oxygen flowing to it. The amount ofwall oxygen being used was calculated for all mechani-cally ventilated patients. We could not account for morethan 240,000 gaseous liters (74 liquid gallons) per day ofoxygen consumption.

There are numerous possible causes for this potentiallyavoidable gas loss. Wall outlets should be regularly in-spected to ensure they do not leak. Unused flow metersshould be monitored to ensure they are shut off. Locationsfor such unused flow meters include clinics, radiologyfacilities, cardiac catheterization units, postoperative re-covery units, and operating rooms. Oxygen flow to manualresuscitators should be turned on only when needed for

Table 2. Oxygen Utilization Survey of Beth Israel Deaconess Medical Center*

Type of Use/LocationNumberof Beds

Percent ofPatients on O2

Number ofPatients on O2

Average Flow(L/min)

Total DailyUse (L/d)

Total DailyUse (gallons/d)

Floors 320 30 96 4 552,960 170Intensive care unit 65 70 49 4.5 315,900 97Operating room 18 100 18 2 51,840 16Postoperative recovery 23 100 23 4 132,480 41Cardiac catheter suite 6 100 6 6 25,920 8Cardiac catheter holding 16 100 16 4 69,120 21Emergency department 51 35 18 4 102,816 32Total accounted usage 1,251,036 384Unaccounted usage 240,969 74

*Data obtained during a 1-day survey.



patient ventilation, and then promptly returned to the offposition. Each manual resuscitator, if left running at 15 L/min, will consume 21,600 L or 6.6 liquid gallons per day.Once inappropriate oxygen use is reduced to a minimum,other opportunities for conservation should be considered.

Oxygen Conserving Options

Reservoir cannulas have been available for years, butare not commonly used in hospitals because of their cost.There are 2 types of device currently available: the Oxy-mizer Reservoir Cannula and the Oxymizer Pendent Can-nula (Chad Therapeutics, Chatsworth, California). The res-ervoir cannula places the reservoir directly below the nose,whereas the pendent model moves the reservoir below thechin. Both models operate without pneumatic or electronicflow controls. When compared to standard oxygen cannu-las, these devices have been shown to sustain a givenoxygen saturation at rest while using 50–75% less oxy-gen.6-10 If we deployed these devices on half of our pa-tients who were receiving oxygen via cannula in our WestCampus facility, we estimate we could conserve over 75liquid gallons of oxygen per day. This would representover 15% of our daily consumption. Cylinder life wouldlikewise be extended. Though the cost of these disposabledevices does not merit their regular hospital use, they couldbe stored and deployed during mass casualty events whenoxygen conservation is critical. We calculated that if weconverted half of our nasal cannula patients to oxygen-conserving cannulas we would reduce our bulk systemconsumption by 20–50 liquid gallons per day. Similarly,they could extend the life of our cylinder oxygen supplies.Electronic or pneumatic flow controllers that deliver ox-ygen only during inspiration also conserve oxygen but aremore expensive. Electronic conservers require batteriesand therefore require substantial in-storage maintenance.Reservoir cannulas offer the more economical and simpleoption.

Another method to extend oxygen resources would beto convert patients with artificial airways from continuousaerosols to heat-and-moisture exchangers and supply low-flow oxygen to sustain a targeted oxygen saturation. Astandard continuous aerosol will generally be powered by12–15 L/min of oxygen. A flow of 15 L/min of gaseousoxygen running for 24 hours will require 6.6 liquid gallonsof oxygen. As a last resort it may be prudent to targetlower patient oxygenation levels. By reducing oxygen flowsand accepting oxygen saturations of 90–92% instead of� 95%, both cylinder and bulk oxygen resources could beextended.

Summary

Whatever emergency contingency an institution decidesto adopt, one thing is clear: when a disaster happens, if you

don’t have what you need, it will not be available quickly,or probably not at all. There is a limit to the amount ofcylinder oxygen that can be stockpiled in preparation for amass casualty event, and those stockpiles will be limited intheir duration. Some amount of cylinder gas should beavailable for surge situations, as long as available storagecomplies with state and federal regulations. Retrieval, dis-infection, and exchange of cylinders must be facilitated inorder to ensure resupply. Clear signage must be in place sothat staff not familiar with oxygen storage locations areable to quickly locate those supplies.

Personal oxygen concentrators and portable LOX sys-tems have limited value but can be used to support se-lected areas. Larger-capacity concentrators and LOX sys-tems are attractive options for developing oxygen deliverysystems at alternative care sites or to support bulk deliverysystems. They must be accompanied by manifold systemsthat match the total system output and provide a mecha-nism that regulates the flow to the patient. If mechanicalventilation is anticipated in these locations, they must adaptto the requirements of the ventilators that are available foruse.

Routine maintenance should be performed on com-pressed-gas systems to eliminate system leaks that causeunnoticed but substantial system consumption. Organizedefforts must be developed to reduce the volume of oxygenconsumed in institutions during mass casualty events. Staffmust ensure that unnecessary use of oxygen is curtailed.Elective surgical procedures should be postponed.

Bulk oxygen delivery areas must be accessible and clearlymarked. Blockage of access areas can prevent timely re-supply of critical commodities. Hospital staff should betrained and certified in the process of refilling bulk oxygensystems in case vendor personnel are not available to per-form this duty. Regular monitoring of system levels mustbe performed.

Dwight D Eisenhower once said, “In preparing for bat-tle I have always found that plans are useless, but planningis indispensable.” One cannot anticipate all the needs andcomplications of a mass casualty event, but some fore-thought can lessen their impact on how we respond to theirchallenges.

REFERENCES

1. FDA and NIOSH public health notification: oxygen regulator firesresulting from incorrect use of CGA 870 seals. http://www.fda.gov/cdrh/safety/042406-o2fires.html. Accessed November 29, 2007.

2. National Fire Protection Association. NFPA 99 standard for healthcare facilities 2002 edition. April 30, 2002

3. Freisen RM, Raber MB, Reimer DD. Oxygen concentrators: a pri-mary oxygen supply source. Can J Anesth 1999;46(12):1185-1190.

4. Taylor-Wharton Gas Equipment Division. VT-21-250 psi verticalstorage tanks: 1,500 gallons through 15,000 gallons. September 3,2002. http://www.taylor-wharton.com/pages/literature/tech%20man%20pdf/bulk/bt_478.pdf. Accessed November 29, 2007.



5. Stoller JK, Stefanak M, Orens D, Burkhart J. The hospital oxygensupply: an “O2K” problem. Respir Care 2000;45(3):300-305.

6. Carter R, Williams JS, Berry J, Peavler M, Griner D, Tiep B. Eval-uation of a pendent oxygen-conserving nasal cannula during rest.Chest 1986;89(6):806-810.

7. Tiep B, Nicotra B, Carter R, Belman MJ, Mittman C. Evaluation ofa low-flow oxygen conserving nasal cannula. Am Rev Respir Dis1984;132(3):500-502.

8. Tiep BL, Nicotra B, Carter R, Phillips R, Otsap B. Evaluation of anoxygen-conserving nasal cannula. Respir Care 1985;30(1):19-25.

9. Tiep BL, Belman MJ, Mittman C, Phillips R, Otsap B. A new pendentstorage oxygen-storage nasal cannula. Chest 1985;87(3):381-383.

10. Soffer M, Tashkin DP, Shapiro BJ, Littner M, Harvey E, Farr S.Conservation of oxygen supply using a reservoir nasal cannula inhypoxemic patients at rest and during exercise. Chest 1985;88(5):663-668.

Discussion

Hanley: How much oxygen will weneed in a mass casualty event? Yousaid you currently use 1.5 million li-ters per week. I calculated it quicklywithout a calculator, so I may bewrong, but I think that is 150 liters perminute, which is 15 patients on 10liters per minute. That’s not that muchoxygen.

Ritz: In a pandemic environment youwould move everybody who is not onoxygen out of the hospital if you can,and convert non-ICU hospital bedsinto ICU beds. We made some calcu-lations about where the oxygen goesin our hospital. The ORs [operatingrooms] don’t use that much under nor-mal circumstances. PACUs [post-an-esthesia care units] use much moreoxygen. Part of the culture in recov-ery units is that the manual resuscita-tors run continuously, so the gas con-sumption is pretty high.

We found that in our ICUs, becausewe use in-line suction catheters andtherefore don’t take the patient off theventilator, we don’t have manual re-suscitators running continuously. NI-CUs [neonatal intensive care units]tend to use a lot of gas because theyhave a lot of their manual resuscita-tion devices always running. And theirventilators are little less economicalbecause they often use continuousflows, particularly oscillators. We haddifficulty estimating the oxygen useper patient.

Out on the floors we estimated thatwe had about 100 patients out of 350on low-flow oxygen at any one pointin time. Just doing some quick sur-

veys, we usually found that those ox-ygen flows were 2 to 4 liters perminute. The floors didn’t consume alot of oxygen. The ICUs used a mod-est amount, the NICUs a lot, ORs abit, PACUs a lot. But it’s difficult tomake these kinds of estimates.

We didn’t account for all of the leak-age in the system and the various labsand clinics and other places, althoughthe clinics tended to use cylinder ox-ygen instead of wall oxygen. It’s dif-ficult to get a clean handle on oxygenconsumption.

O’Laughlin: I’ve looked at settingup oxygen at an alternate care site andI found logistical nightmares associ-ated with that. I’m talking about anarea that doesn’t have built-in oxygentubing. Strictly setting up from scratch,this is based on the original oxygen-delivery layout that came out of Den-ver for alternate care sites and talkingabout oxygen delivery at those sites.1

They talked about liquid oxygen setup.I did some numbers for a 50-patient

unit using sets of 2 H-size tanks (per4–5 patients) combined with a mani-fold such as the Minilator with a fixedO2 rate, using Oxymizer nasal cannu-las, and I still would have to deal withcomplete turnover of the tanks twicein 24 hours. The logistics of cylin-dered oxygen at an alternate care siteseem too complex for cylinders to re-ally be useful. What are your thoughtsabout the logistics and costs of settingsomething up?

1. Lam C, Waldhorn R, Toner E, InglesbyTV, O’Toole T. The prospect of using al-ternative medical care facilities in an influ-enza pandemic. Biosecur Bioterror 2006;4(4):384-390.

Ritz: The most common, most avail-able cylinder is an E-size cylinder.There are lots of those. There are notnearly as many larger cylinders formedical use in most hospitals. How-ever, there is a substantial amount ofcylinder oxygen in larger-volume cyl-inders sized for industrial use, andthere is industrial oxygen available.

The problem is coming up with thehardware to distribute this oxygencompletely, to various patient caresites. With enough preplanning andsupport from the vendor you could doit, but it’s sort of like having to buyinsurance from them. You’d have tosay, “If I do this, can you guaranteethat you can ship this much this fast?”

The manifolds are not hard to makeor buy. There’s a lot of variations ofthat kind of equipment. The questionis whether you can get the cylindersand do you have somebody aroundwho can change those cylinders out?I’m skeptical that you can get enoughcylinders for a long enough period oftime in the scenarios we’re consider-ing.

O’Laughlin: I talked to a major ox-ygen manufacturer, and they saidthey’d be thrilled to store cylindersthat we would purchase up front; theywould house them for us, for a verynice fee. That’s an option, but is it theright balance of resources and the rightway to spend the money?

You mentioned the question of us-ing manufacturing (nonmedical) oxy-gen supplies. I think that the grade ofoxygen isn’t necessarily consideredthe same, but is there the potential touse industrial grade oxygen in a med-ical system? Are there majorequip



ment needs to try to hook the systemup, or is it relatively the same, or is itnot considered appropriate at all?

Ritz: All the oxygen comes from thesame plants, whether it is industrial ox-ygen or medical oxygen. I think the cyl-inder connections are the same also.There is an indexing system [DISS (di-ameter-index safety system)] that re-quires the cylinder threads to be of cer-tain diameter and configuration. So theuse of industrial oxygen is not a partic-ular impediment. That seems like a rea-sonable thing to do. There are labelingregulations and CGA [Compressed GasAssociation] regulations that prevent usfrom doing it currently, but there’s nodifference between the gases.

Rubinson: With regard to alternatecare facilities, the 3 oxygen optionsare compressed gas, gas production on-site, and liquid oxygen. The portableon-site oxygen-production units I haveseen provide approximately 33 liters aminute for $65,000, so that’s not agreat option for most communities. In-dividual concentrators can be used, butthey require electricity running to eachpatient site, although at some sites thatcould work. They would all have largeelectricity requirements to produce ox-ygen for hundreds of patients, whichis the number being considered forlarger alternate care facilities. For liq-uid oxygen we calculated for 200 to250 patients on an average of 4 litersa minute, which is a total of 1,000liters per minute. So we addressed flowin terms of having sufficient vapor-izer capability and also having totaloxygen capacity.

We worked with one large vendorandlookedatMicro-BulkPermaCylcyl-inders. Because the Micro-Bulk distri-bution truck in our area is different fromthe bulk trucks, we thought this was ap-pealing because our alternate care facil-ity would not be competing with thehospitals for LOX distribution. We wereconcerned about how quickly theywould have to resupply, and it’d be be-tween 1 and 2 days.

So now we’re looking at the optionof the truck that they use when a hos-pital is just about to introduce a LOXsystem or when they need a temporaryalternative external system. Unfortu-nately, there are actually very few ofthose in the country. But we are work-ing at being able to adopt this system inSeattle/King County. So there are somepotential solutions. I think it’s probablygoing to cost between $100,000 and$200,000 per 200 to 250 patients in al-ternate care facilities.

With regard to manufacturing oxy-gen, the key issues are that the electricalgrid has to be up and that the workerscan make it to work at the oxygen plant.Distribution is also a concern. There are5 or 6 major oxygen manufacturers, andthey have limited distribution redun-dancy, so if multiple regions are con-currently impacted by an epidemic andthe workers are less available to manu-facture and distribute it, LOX could bedelayed or unavailable. There are thou-sands of compressed-gas distributors,and about 15% of their business is med-ical oxygen. There may be the possibil-ity of the FDA [U.S. Food and DrugAdministration] relaxing regulations ifnonmedical cylinders need to be used,but we can’t rely on that. Even if regu-lations are relaxed, getting enough com-pressed gas to replace LOX for hospi-tals will be challenging.

Daugherty: How much oxygenmight be saved by canceling electivesurgeries? How many days of oxygenwould that buy? Lewis said that hos-pitals need to plan to be independentfor the first 10 days of an emergencymass critical care event.

Ritz: With LOX systems the intentis that you never get into your reservesupply. The only loss from your re-serve system is the natural evapora-tion that occurs. LOX cylinders arefilled only half full or slightly less thanhalf full, but it seems like there’s afair amount of reserve in a hospitalliquid oxygen system. We found thatthe NICU and PACU had the highest

oxygen consumption. In an OR theanesthesia machine consumes oxygen,but there aren’t a lot of other sourcesof gas consumption in an OR; it’s justthat there’s a lot of them and they runa long time. The oxygen flow goingthrough an anesthesia machine can bequite economical. The ORs’ consump-tion might not be so great that if youstopped it, it would extend your oxy-gen supply by 3 or 4 days.

Branson: Isn’t it the practice for allanesthesia machines to have a low flowof oxygen 24 hours a day, to go throughthe CO2 absorber to prevent the de-velopment of carbon monoxide? Couldturning off that flow conserve an im-portant amount?

Talmor: There’s a very low flowthrough the machine; it’s less than a li-ter per minute. Anesthetics are typicallydelivered with low flows these days, soabout a liter of oxygen per minute, ex-cept at the end of the anesthetic, whenit’s turned up to remove the inhaledagents, and then the flow goes up toabout 15 liters per minute. The substan-tial waste in the OR is if the oxygenflow is not turned back down. If youhave 35 ORs, like we do, that could addup to substantial waste.

Hanley: Are there legal restraintsthat prevent oxygen suppliers fromgouging hospitals if we suddenly havean increase in demand?

Rubinson: I don’t know if there arelegal restraints, but I can tell you thatthe vendors came through after [hurri-cane] Katrina. The major manufacturersare under strict antitrust rules becausetherearesofewmanufacturers. I’vebeentold that if a plant goes down, the othercompanies will often fill that compa-ny’s obligations. So I think, typically,the oxygen industry has done the rightthing when needed, but I don’t know ifthere is anything specific to oxygen thatwould get them at gouging any morethan any other medical equipment oranything else that’s in short supply.



oxygen supplies during a mass casualty situationoxygen supplies when the demand is suddenly...

Documents