What Is the Recommended Pulse Oximetry SpO2 Range to Maintain Using Supplemental Oxygen in a Post-Cardiac Arrest Patient: The 90–98% Target With a 92% Floor

The recommended range is 90% to 98%, with most protocols treating 92% as the working floor and 98% as the ceiling for patients who need supplemental oxygen after return of spontaneous circulation. The American Heart Association's post-arrest care guidance frames this as avoiding both hypoxemia and hyperoxia, so the practical answer is titration to the lowest inspired oxygen that holds saturation in that band, not a fixed high-flow setting.

Why 90–98% Became the Target Instead of "the Higher the Better"

The shift matters because oxygen stopped being treated as a harmless drug. For decades, after a heart stopped and restarted, high-flow oxygen felt protective, and saturation above 98% was rarely questioned. Post-arrest care guidance from the American Heart Association now explicitly warns against both hypoxemia and hyperoxia, with the recommended saturation band sitting between 90% and 98% for patients on supplemental oxygen. That range exists because two failure modes pull in opposite directions.

On the low side, below roughly 90%, hemoglobin saturation falls off the steep part of the oxygen dissociation curve. A saturation of 88% corresponds to a arterial oxygen tension near 55 mmHg in a normal-acidosis patient, and small further drops produce large falls in delivered oxygen. Post-arrest brains are already vulnerable after global ischemia, so a saturation dipping into the 80s during transport or a transfer between units is a genuine event, not a monitor artifact to shrug off.

On the high side, sustained hyperoxia above 98–100% is associated with worse outcomes in observational post-arrest cohorts, through mechanisms that include coronary vasoconstriction, reperfusion injury from reactive oxygen species, and absorption atelectasis. The evidence is not a clean randomized trial showing mortality benefit at every threshold, and honest protocols say so. What the guidance gives clinicians is a defensible operating band rather than a target number.

The boundary conditions are where people get into trouble. A patient with chronic obstructive pulmonary disease and a documented CO2-retaining phenotype may have a personal baseline saturation of 88–92%, and pushing them to 96% can drive carbon dioxide retention. A patient with severe anemia may read 97% while carrying inadequate oxygen content. Pulse oximetry measures saturation, not content and not delivery. The 90–98% target is a starting frame, not a treatment endpoint.

The practical translation: prescribe a target range in the orders, not a flow rate. "Titrate oxygen to keep SpO2 at 92–96%" tells the bedside nurse or paramedic what to do. "Oxygen 15 liters by non-rebreather" tells them nothing about when to stop.

The 92% Floor and 98% Ceiling: How Each Number Was Set

The 92% floor and 98% ceiling come from different logic, and treating them as symmetrical is a mistake. The low limit is a safety floor tied to the dissociation curve and the risk of rapid desaturation when a patient's oxygen consumption rises, as it does during shivering, fever, or a seizure. The high limit is a ceiling tied to evidence that supranormal saturation offers no benefit and carries plausible harm. European Resuscitation Council guidance and the AHA both land in the same neighborhood without prescribing an identical number, which is why you will see 90–98%, 92–98%, and 94–98% in different protocols.

Post-arrest patients on mechanical ventilation are a special case. The ventilator delivers a set fraction of inspired oxygen, and the ICU team titrates that fraction down as the patient stabilizes, aiming for the lowest fraction that holds saturation above the floor. A patient who arrives from the field on 100% oxygen should not stay there by default. Weaning the fraction down toward 0.30–0.40 while watching saturation is standard practice in most post-arrest bundles, and the number to watch is the fraction, not the liter flow, once the patient is intubated.

The ceiling of 98% has an important exception that rarely gets mentioned in patient-facing material. In a patient with severe anemia, carbon monoxide exposure, or a right-to-left shunt, the saturation reading can sit comfortably at 97% while tissue oxygen delivery is inadequate. Saturation is a ratio, not a quantity. A patient with a hemoglobin of 6 g/dL can have a normal saturation and still be in oxygen debt. Clinicians who treat 98% as a guarantee of adequate oxygenation are misreading the test.

The floor also has a documented failure mode. In a patient with dark skin pigmentation, pulse oximetry has a well-described tendency to read falsely high at low saturations, and the error grows as saturation falls. The US Food and Drug Administration has held advisory committee meetings on this exact issue, and the discrepancy is large enough that some studies report occult hypoxemia in roughly 1 in 5 to 1 in 3 Black patients whose readings looked acceptable. A saturation of 92% on the monitor may correspond to a true arterial saturation several points lower in some patients. That doesn't change the target band, but it changes how much trust to place in a single reading at the floor.

Reading the Monitor Correctly Before You Adjust Anything

Before adjusting oxygen based on a saturation number, confirm the number is real. Pulse oximetry depends on a pulsatile signal, and the common failure modes are mechanical, not physiological. Poor perfusion from hypotension, cold extremities, or vasopressor use can produce a weak signal and a falsely low or erratic reading. Motion artifact from shivering or a patient being transported produces dropouts. Nail polish, artificial nails, and dirt on the sensor site attenuate the signal. Each of these can trigger an unnecessary oxygen increase in a patient who is already at target.

Check the plethysmographic waveform, not just the number. A good waveform has a clear pulsatile contour with each beat, and the displayed saturation usually tracks with it. A flat or chaotic waveform means the number cannot be trusted. This is where the failure case lives: a patient with a saturation of 87% on a monitor with a poor waveform who gets oxygen increased to 100% and is held there for an hour while the actual arterial blood gas shows a saturation of 96% on room air. The oxygen was never needed, and the brief hyperoxia was avoidable.

Site selection matters more than most people think. The finger is the default, but in a patient with poor peripheral perfusion, the ear lobe or forehead sensor can produce a more reliable signal because those sites are less affected by vasoconstriction and motion. Forehead sensors in particular have been studied for their ability to detect central hypoxemia faster than finger probes, though they are not universally available and are more prone to ambient light interference.

If the reading is inconsistent with the clinical picture, get an arterial blood gas. A saturation reading and a blood gas drawn at the same moment are the gold standard comparison, and the blood gas also gives the arterial oxygen tension and the acid-base status that the saturation alone cannot. In a post-arrest patient who is not stabilizing, the blood gas is not optional. It is the test that tells you whether the monitor is lying.

Oxygen Delivery Devices and What Each One Actually Achieves

The device matters because the target range determines how much oxygen to deliver, and different devices deliver very different amounts at the same liter flow. A nasal cannula at 1–6 liters per minute delivers an inspired oxygen fraction of roughly 24–44%, but the actual fraction depends on the patient's breathing pattern and the entrainment of room air. A simple face mask at 5–10 liters delivers roughly 35–55%. A non-rebreather at 10–15 liters delivers 60–95%, and a high-flow nasal cannula can deliver a precise fraction at high flow with humidification. The same patient can be under-oxygenated on one device and over-oxygenated on another at the same liter flow.

The practical rule: if a patient is at target on a non-rebreather, the next move is usually to step down to a lower-delivery device while watching the saturation, not to keep the non-rebreather running because it feels safer. Step-down decisions should be documented with the saturation before and after the change. A patient who drops below the floor after stepping down goes back up a level and gets reassessed.

The failure mode here is the "set it and forget it" non-rebreather. A patient transported on 15 liters by non-rebreather arrives at the emergency department and the device stays on for the next two hours because no one was assigned to reassess. That produces sustained hyperoxia in a patient who may have been at target on 2 liters by nasal cannula. The fix is a documented reassessment interval, usually every 15–30 minutes during the unstable phase and less often once stable.

Titration in Practice: A Step-by-Step Adjustment Sequence

The order of operations matters, and doing it out of order produces either unnecessary oxygen or unnecessary desaturation. The sequence below reflects how most post-arrest bundles structure the first hour after return of spontaneous circulation, and it assumes the patient is being monitored continuously with a reliable waveform.

1: Confirm the sensor site has a good plethysmographic waveform and a reading consistent with the clinical picture. If not, reposition the sensor, try an alternative site, or obtain an arterial blood gas before acting on the number. 2: Set an initial inspired oxygen fraction or flow that is high enough to prevent desaturation during the first minutes, typically 100% in the field or on arrival, then move immediately to titration. 3: Titrate down in steps, watching the saturation for 30–60 seconds at each step, until the saturation sits in the 92–96% zone rather than at the ceiling. A patient at 100% is at the ceiling, not at target. 4: Recheck after every change, by device and by patient position, and document the saturation before and after. If the patient desaturates below the floor, step back up one level and reassess for a cause rather than escalating to maximum. 5: Once stable, set a reassessment interval and a standing order that specifies the target range, not a fixed flow. The order should say "titrate to keep SpO2 92–96%" and name who is responsible for the reassessment.

The failure-handling close: if titration produces repeated desaturation below the floor despite escalating support, the problem is usually not the oxygen setting but the patient's work of breathing, airway, or hemodynamics. Look for a reversible cause before pushing the fraction back to 100%, and get a blood gas to separate a monitor problem from a physiology problem.

Special Populations Where the Standard Range Needs Adjusting

Three groups deserve their own frame because the standard range can mislead. The first is the chronic CO2 retainer with COPD or obesity hypoventilation. For these patients, the target is usually individualized to their documented baseline, often 88–92%, with the goal of avoiding both hypoxemia and carbon dioxide retention. Pushing them to 96% can produce a respiratory acidosis that the post-arrest brain tolerates poorly. The target range should be written in the chart with the baseline documented, not assumed from the general band.

The second group is patients with severe anemia or active bleeding. Saturation can read normally while oxygen content is inadequate, so the target range is necessary but not sufficient. The relevant numbers become hemoglobin and transfusion thresholds, and the saturation band is a floor to defend rather than a treatment target.

The third group is patients with dark skin pigmentation, where pulse oximetry has a documented tendency to read falsely high at low saturations. The FDA has convened advisory committee meetings on this disparity, and clinical studies have reported occult hypoxemia in a meaningful fraction of Black patients whose readings appeared acceptable. The practical response is a lower threshold for confirming with arterial blood gas and a lower threshold for using an alternative site or sensor, not a change in the target band itself.

The boundary condition that ties these together: the 90–98% range is a population-level recommendation, and individual targets should be documented with the reason. A protocol that says "keep everyone at 94%" is easier to write and worse to follow than one that says "keep most patients at 92–96%, keep documented CO2 retainers at their baseline band, and confirm with blood gas when the reading and the patient disagree."

Common Mistakes That Push Patients Outside the Band

The most common error is failing to wean. A patient arrives on 100% oxygen, stabilizes, and stays on 100% because no one wrote an order to titrate down. The fix is a standing order that specifies the target range and a responsibility assignment, not a hope that someone will remember.

The second error is treating the saturation number as a treatment endpoint rather than a monitor. Increasing oxygen to chase a number from 91% to 94% while ignoring a rising respiratory rate or falling blood pressure addresses the display and not the patient. The saturation is one signal among several, and the post-arrest patient's trajectory is read from the whole picture.

The third error is over-trusting a single reading in a patient with poor perfusion. Vasopressor use, cold extremities, and hypotension all degrade the pulse oximetry signal, and the response should be to confirm, not to escalate. A blood gas is cheap compared with an hour of unnecessary hyperoxia.

The fourth error is failing to document the device and the flow alongside the saturation. A note that says "SpO2 95%" without saying "on 4 liters nasal cannula" cannot be used to track the trajectory or to hand off. Write the device, the flow or fraction, the site, and the waveform quality.

Costs and Workflow: What the Target Range Saves and What It Costs

The cost calculation is not about the oxygen itself, which is cheap, but about the monitoring and reassessment time the target range requires. A typical post-arrest patient in an intensive care unit gets continuous pulse oximetry, intermittent arterial blood gases, and a nursing reassessment every 15–30 minutes during the unstable phase. The blood gas is the expensive item, and its use is what separates a protocol that confirms suspicious readings from one that chases numbers.

A worked example: a patient with three blood gases in the first six hours versus one. The difference is two additional arterial punctures, two additional laboratory runs, and the nursing time to perform them. That cost is justified when the readings are inconsistent or the patient is not stabilizing, and it is waste when it is done on a schedule regardless of the clinical picture. The target range tells you when to confirm, not how often to sample by habit.

The workflow cost of the target range is the reassessment itself. A protocol that specifies 92–96% creates a nursing task at every handoff and every device change, and that task is the mechanism by which hyperoxia is prevented. A protocol that specifies a fixed flow creates no such task and depends on individual vigilance. The evidence for the range is strongest at the level of the bundle, not the individual number, and the bundle includes the reassessment.

FAQ

Is 100% oxygen ever appropriate after cardiac arrest?

Yes, for the initial resuscitation and transport phase, when the priority is preventing any desaturation during an unstable period. The AHA guidance supports starting with high inspired oxygen and then titrating down once return of spontaneous circulation is established and the patient is being monitored. The error is staying at 100% after the patient stabilizes. A saturation of 100% on 15 liters by non-rebreather is at the ceiling, not at target, and the next action is a step-down with reassessment.

What if the saturation reads 89% but the patient looks fine?

Treat the number as a signal to confirm, not a signal to ignore. A reading below the floor in a patient who appears well is either a real desaturation that the clinical picture has not yet caught up with or a monitor artifact from poor perfusion, motion, or site selection. Check the waveform, reposition the sensor, try an alternative site, and obtain an arterial blood gas if the discrepancy persists. In a post-arrest patient, the cost of confirming is low and the cost of missing a real desaturation is high. Document the confirmation either way.

Does the target range differ for a patient with COPD?

Yes, and the difference should be documented rather than assumed. A patient with a documented CO2-retaining phenotype often has a baseline saturation in the 88–92% range, and targeting 96% can produce carbon dioxide retention and respiratory acidosis. The general band still applies as a safety frame, but the individual target should be set at the patient's documented baseline and written into the orders with the reason. Confirm the baseline from prior records rather than guessing at the bedside.

How often should oxygen be reassessed after cardiac arrest?

Every 15–30 minutes during the unstable phase, and after every device change, position change, or transfer, until the patient is stable and the target range is being held on a stable device. Once stable, the interval can lengthen, but the standing order should specify the target range and the reassessment responsibility rather than a fixed flow. The reassessment interval is part of the protocol, not an optional extra.

Can pulse oximetry be trusted in patients with dark skin?

It can be used, but with a lower threshold for confirmation. The FDA has reviewed evidence that pulse oximetry can read falsely high at low saturations in patients with dark skin pigmentation, and clinical studies have reported occult hypoxemia in a meaningful fraction of Black patients whose readings appeared acceptable. The practical response is to confirm suspicious or borderline readings with arterial blood gas and to consider alternative sites or sensors. The target band does not change; the confidence in a single reading does.

What is the difference between saturation and oxygen content?

Saturation is the percentage of hemoglobin binding sites occupied by oxygen, while oxygen content is the total amount of oxygen carried in the blood, including the dissolved fraction and the hemoglobin concentration. A patient with severe anemia can have a normal saturation and inadequate oxygen content, which is why the saturation band is a floor to defend rather than a complete treatment target. Hemoglobin and transfusion thresholds become relevant in those patients.

Is there a role for high-flow nasal cannula in post-arrest care?

Yes, in patients who need precise titration and support for work of breathing without immediate intubation. High-flow nasal cannula delivers a titratable inspired oxygen fraction at high flow with humidification, which allows finer control than a non-rebreather and can reduce the work of breathing. The target range still applies, and the fraction should be titrated down as the patient stabilizes. Device selection should follow the patient's trajectory, not a default.

What should be documented at every reassessment?

The saturation, the device, the flow or inspired fraction, the sensor site, the waveform quality, and the patient's clinical trajectory. A note that records only the saturation cannot be used to track the response to titration or to hand off to the next clinician. Documentation of the device and the flow is what makes the target range auditable and what prevents the "set it and forget it" failure.

This article is general educational information about oxygen targets after cardiac arrest and is not a diagnosis, treatment plan, or substitute for the judgment of a qualified clinician. Individual targets depend on the patient's baseline, comorbidities, and current physiology, and should be set by the treating team with local protocols and specialist input. For immediate danger or a patient who is deteriorating, contact local emergency services rather than relying on any general guidance.