Field Brief: Cool First, Transport Second — The Phoenix Cold Water Immersion Study & the New Science of Heat Emergencies
It is 114°F on a Phoenix sidewalk in July. Engine 21 and Rescue 18 arrive for a 58-year-old man found unresponsive beside a bus stop. He is supine on the concrete, no shade, skin hot to the touch. Rectal temperature: 107.1°F. GCS 6. The closest hospital is twelve minutes away on paper — but this is Phoenix in July, and traffic is already stacked up.
The old algorithm says load him in the rig, crank the A/C, start two large-bore IVs, hang fluids, and drive. The new algorithm — the one Phoenix Fire ran on 183 patients in a study just published in Prehospital Emergency Care — says you put him in a body bag, fill it chest-high with ice water, and start cooling on scene. Twelve minutes later he arrives at the ED at 102°F with a GCS of 12.
That difference between the old and new is what this post is all about.
What "Fighting Fire with Ice" Actually Reported
In March 2026, Hosseinipour and colleagues published Fighting Fire with Ice: A Multisite Collaboration to Evaluate the Impact of Prehospital Cold Water Immersion on Heat Stroke Patients in Prehospital Emergency Care. The paper describes Phoenix Fire Department's implementation of a novel prehospital cold-water immersion (CWI) protocol across the entire urban service area, then evaluates the patients who received it.
The protocol is exactly as low-tech as it sounds. Crews carry a heavy-duty plastic enclosure with handles — commercial rapid-cooling bag, repurposed body bag, same idea. When a patient meets criteria — tympanic temperature at or above 104°F with CNS dysfunction attributable to heat exposure — they place the patient in the bag inside the rig or on scene, fill it chest-high with bagged ice and water, and start cooling immediately. Transport begins when crews are loaded. Cooling continues en route. They pull the patient off the rig at the ED ramp with the bag still slung underneath, hand off to ED staff, and the receiving team takes over.
The study population: 183 patients across multiple agencies who met the inclusion criteria of tympanic temperature ≥40°C plus associated CNS dysfunction. The headline numbers:
72.1% of patients had a temperature decrease between prehospital recognition and ED arrival, with an average drop of 3.6 ± 2.7°F.
Median Glasgow Coma Scale climbed from 8 (IQR 4–10) prehospital → 9 (IQR 3–14) at ED arrival → 15 (IQR 14–15) at discharge.
The Phoenix Fire Department alone used the protocol more than 450 times since spring 2024 by the study cutoff.
The authors' conclusion is the part you should forward to your medical director: a prehospital CWI protocol is feasible in a large urban EMS system, provides rapid temperature reduction, and produces improved neurologic status in patients with heat stroke. Feasible. In a city of 1.6 million. On a fire-based ALS service. With existing personnel and a few hundred dollars of plastic per rig.
That feasibility finding is the part that buries the dogma.
Why Load-and-Go Kills Heat Stroke Patients
Heat stroke is one of the few prehospital emergencies where the rate of treatment, not the speed of transport, is the only thing that matters to outcome. The mortality curve isn't time-to-hospital. It's time-above-40°C. Every minute over the threshold compounds endothelial injury, mitochondrial dysfunction, and the inflammatory cascade that ends in DIC, multi-organ failure, and either death or permanent neurologic damage.
Belval and Casa's cohort data, synthesized in the NAEMSP 2022 consensus statement on prehospital exertional heat stroke, gives us the threshold the field operates on. A cooling rate at or above 0.15°C per minute (roughly 0.27°F per minute) was associated with a 0% case fatality rate. Below that rate, case fatality climbed to 4.41%. Complication rates were 0.77% in adequately cooled patients vs. 22.46% in the under-cooled group. The signal is not subtle.
The 2024 SCCM heat stroke guideline (Sweeney et al., Critical Care Medicine) translates the signal into a strong recommendation: active cooling over passive cooling, and prioritize the modality that achieves the fastest cooling rate — which is ice- or cold-water immersion at 1–12°C. The target is to reach the cooling endpoint within 30 minutes of heat stroke recognition.
Now run the math on the heat stroke patient your service ran last August. The crew drove 18 minutes with A/C blasting, two liters of room-temp normal saline running wide open, a sheet wetted with tap water draped across the chest, and cold packs in the axillae and groin. What was the cooling rate?
If the answer is "we didn't measure it," that puts the service in the same bucket as most of US ground EMS, where a 2021 review found roughly 60% of state EMS treatment protocols do not include CWI as an option for heat stroke. The 2024 Annals of Emergency Medicine implementation paper from the same Phoenix group documented that the wetted-sheet-and-fan approach written into many civilian protocols achieves a cooling rate well below the 0.15°C/min threshold. Cold packs to the groin and axillae alone measured at roughly 0.03–0.07°C/min in older controlled studies. CWI ranges from 0.15 to 0.35°C/min depending on water temperature.
The standard answer most of EMS gives a heat stroke patient is, by measurement, three to five times too slow to prevent organ damage.
That isn't a sentiment. It's a rate.
The Two Heat Emergencies Your Protocols Probably Confuse
Before the operational specifics, name the patient correctly. Heat stroke comes in two flavors, and civilian EMS often blurs them.
Exertional Heat Stroke (EHS) is the high-school football player on the August two-a-day, the firefighter on a working structure fire, the Marine on a 14-mile road march. Young, usually healthy, sweating profusely, collapsing during heat-producing activity. Mortality with prompt cooling: under 5%. Without prompt cooling death-rates climbs fast.
Classic (non-exertional) Heat Stroke (CHS) is the 78-year-old woman found in a third-floor apartment with no A/C during a heat dome. The unhoused patient at the bus stop. The chronic schizophrenic on an anticholinergic regimen who can't sweat. CHS is the dehydrated, dry-skinned, often elderly patient with comorbidities. Mortality in severe CHS is reported at up to 65% even with treatment.
The Phoenix cohort included both presentations. The pathophysiology converges — once core temperature crosses 40°C/104°F with CNS dysfunction, the immediate treatment is the same: aggressive cooling, now, on scene. The Phoenix data refutes the older argument that classic-pattern patients are "too fragile" for CWI. They tolerated it. They cooled. Their neurologic status improved.
The diagnostic move is to confirm core temperature, not skin temperature. Tympanic for initial screening because it's fast and clean. Rectal probe for continuous monitoring during cooling because everything else lies above the 40°C/104°F threshold.
What Phoenix Actually Carries
The kit is intentionally austere. The components in a typical Phoenix Fire heat response:
A heavy-duty, watertight cooling bag with handles. Commercial Ready-Heat or a repurposed transport body bag — both work.
40–60 lbs of bagged ice cached on the rig during heat season.
5–10 gallons of water, often pre-staged in collapsible jugs.
A continuous core-temperature monitor — typically a rectal probe with a thermometer that reads 25–45°C/77-113°F.
Standard ALS resuscitation gear. Nothing exotic.
The whole add-on, including a thermometer that reads in the range you actually need, runs under $500 per rig with ongoing ice resupply as the variable cost. Compared to a Lucas device, an X-Series monitor, or a single dose of suzetrigine, the cost argument evaporates.
The operational catch is core temperature measurement, not the cooling itself. Tympanic and oral thermometers are unreliable in heat stroke, and many top out below the temperatures you actually see. Phoenix uses tympanic to screen and rectal probe to monitor cooling. You cannot cool to a target if you cannot measure where the patient is.
The stopping point matters as much as the cooling. NAEMSP consensus is to cease active cooling at a core temperature of 38.6°C (101.5°F). Overshoot is real, clinically meaningful, and the single most common failure mode once an agency adopts aggressive cooling. Cold patients in DIC do worse, not better.
The Operational Objections, Answered
The objections to civilian-EMS CWI are the same ones you'll hear in your next QA meeting. Here is what the evidence says back.
"It'll wet down the rig." Yes. So does urine, vomit, and a hemorrhaging femur. The cooling bags are watertight. They are meant to be discarded at the ED, like a stretcher sheet, not laundered and reused.
“My transport time is five minutes — why bother?” Because cooling rate matters more than distance to the ED. A 5-minute transport at a 0.05°C/min / 0.09°F/min cooling rate delivers only 0.25°C / 0.45°F of cooling. A 5-minute on-scene CWI initiation at 0.20°C/min / 0.36°F/min delivers 1.0°C / 1.8°F of cooling — and the cooling continues in the rig. The variable is the cooling rate. The transport time is noise.
"We don't have ice on the rig." You can cache it during the local heat season. Phoenix stocks ice in coolers as part of the morning truck check from May through September. The logistics solved itself once medical direction wrote the protocol.
"What about hypothermia from overshoot?" Real risk, well-described, prevented by stopping cooling at 38.6°C/101.4°F and using continuous rectal monitoring. The same risk applies to ED-based CWI and is managed the same way.
"Our state protocols don't allow it." That isn't a clinical objection — it's a leadership and medical-direction problem. The fact that 30-plus states still don't include CWI in their protocol library is the topic of another Black Flag piece on EMS protocol lag. For now, if a service runs in any climate where heat stroke is plausible — and that is most of the continental US for at least three months a year — the medical director should be reading this PEC paper before the next heat wave.
What To Do Monday Morning
For Working Medics:
If the service does not currently carry a rapid-cooling option, take the Phoenix paper to medical direction. The DOI is 10.1080/10903127.2026.2636148.
Know the diagnostic criteria for heat stroke: core temperature at or above 40°C /104°F plus CNS dysfunction (altered mental status, seizure, coma) in a heat-exposure context. Not "the patient feels hot." Measure.
Practice rectal probe placement on a manikin before you need it at 1300 in the parking lot of a music festival. If the service doesn't have a probe, the conversation with the supervisor starts there.
Stop using cold packs to the groin and axillae as the primary cooling modality. Not because the packs do nothing — but because alone, they don't move the needle in 30 minutes.
Stop relying on "cool the patient compartment down" as a treatment plan. A/C cools the patient at a rate roughly equivalent to leaving them on the curb in the shade.
For Medical Directors and EMS Leaders:
Add a CWI or rapid-cooling-bag protocol to the treatment guidelines for both exertional and classic heat stroke. Use the SCCM 2024 guideline as the evidence anchor and the Phoenix paper as the feasibility anchor.
Cache ice and watertight cooling bags on rigs during the local heat season. This is a $200–$500 per rig add and a $50/month ice resupply problem. Not a capital ask.
Implement continuous core temperature monitoring as standard for heat stroke patients. Tympanic to screen; rectal probe during cooling.
Write the 38.6°C/101.4°F cooling stop into the protocol. Overshoot kills the credibility of the entire intervention faster than any other failure mode.
The Bigger Picture
The Phoenix study isn't conceptually new. Casa, Belval, the Korey Stringer Institute, and the military medical community have been making this argument for over a decade. What changed is the evidence base for civilian urban EMS at scale. Until 2024, every defense of "load and go for heat stroke" leaned on a single point: yes, CWI is faster, but it's only feasible at football games and military training facilities with pre-staged cooling resources. It can't be done at scale in urban civilian EMS.
That argument is now dead. Phoenix Fire ran it 450 times in eighteen months across a city of 1.6 million people. Scottsdale Fire adopted a parallel protocol in 2025. Maricopa County's heat-related death count dropped from 265 to 156 between the 2024 and 2025 mid-September snapshots — a multi-factorial outcome that includes cooling centers, public messaging, and outreach, but in which the prehospital cooling protocol is one of the named contributors.
The dogma now belongs to the agencies that haven't adopted it. They will need to explain why, and the explanation cannot be "we never thought about it." The evidence is in the journal. The cost is trivial. The training takes an afternoon.
The patient on the sidewalk in 114°F doesn't have time for the next quarterly protocol committee meeting. You either cool them on scene, or you don't.
The Phoenix study is the part where the field found out: when you cool them, they live with their brain.
Sources
Hosseinipour N, et al. Fighting Fire with Ice: A Multisite Collaboration to Evaluate the Impact of Prehospital Cold Water Immersion on Heat Stroke Patients. Prehospital Emergency Care, March 2026. DOI 10.1080/10903127.2026.2636148. PMID 41739962.
Belval LN, Casa DJ, et al. Consensus Statement: Prehospital Care of Exertional Heat Stroke. NAEMSP, 2022.
Sweeney DA, et al. Society of Critical Care Medicine Guideline for the Treatment of Heat Stroke. Critical Care Medicine, executive summary 2025; full guideline 2024.
Hosseinipour N, et al. Heat Stroke Management Updates: A Description of the Development of a Novel In-Emergency Department Cold-Water Immersion Protocol and Guide for Implementation. Annals of Emergency Medicine, 2024.
Belval LN, Casa DJ, et al. Cooling Modality Effectiveness and Mortality Associated With Prehospital Care of Exertional Heat Stroke Casualties. J Emerg Med, 2022.
Brown T, et al. Status of US Emergency Medical Service Protocols Regarding Pre-Transfer Cooling for Exertional Heat Stroke. PMID 34912642, 2021.
Maricopa County Department of Public Health, Heat Surveillance Reports, 2024 and 2025.
Korey Stringer Institute (UConn), exertional heat stroke clinical resources.
Comments