How Much Oxygen Does a BVM Deliver?

A bag-valve-mask (BVM) connected to an oxygen source and fitted with a reservoir bag can deliver close to 100% oxygen to a patient, but the real-world number swings dramatically depending on the device, the technique, and the seal against the face. Without a reservoir, a BVM connected to supplemental oxygen delivers roughly 40–60% oxygen. With a properly functioning reservoir and a flow rate of 15 liters per minute, the best devices push that figure above 90%. Yet testing of 40 commercially available BVM devices found that eight of them failed to deliver above 85% oxygen even under ideal laboratory conditions, and three delivered below 55%.

How the Reservoir Changes the Math

A standard self-inflating BVM bag holds about 1 to 1.6 liters of air. When you connect oxygen tubing to the bag’s inlet without attaching a reservoir, the bag refills partly with oxygen from the tubing and partly with room air drawn in through a one-way valve. Room air is about 21% oxygen, so the resulting mixture is diluted. One neonatal study demonstrated this starkly: oxygen measured at the rear of the bag (near the oxygen inlet) read 100%, but oxygen at the front of the bag (near the patient port) read only 21% at all tested flow rates when no reservoir was attached.¹ The oxygen simply hadn’t filled the bag before the next squeeze.

Attaching a reservoir bag solves this by providing a secondary chamber that fills with pure oxygen between breaths. When the rescuer squeezes the BVM, the bag draws from this oxygen-rich reservoir instead of entraining room air. A comparison of reservoir types found that a 2.5-liter bag reservoir consistently delivered an oxygen fraction of 1.00, meaning nearly pure oxygen reached the patient port.² Corrugated tube reservoirs, which are sometimes used as substitutes, proved more sensitive to changes in ventilation rate and oxygen flow and delivered less reliably.

Bag refill time matters more than people realize. If the rescuer squeezes the bag too quickly and doesn’t let it fully re-expand between breaths, the reservoir doesn’t have time to refill with oxygen. The result is a lower oxygen concentration on the next breath even when the equipment is set up correctly. This is one reason guidelines emphasize a steady, deliberate squeeze-release rhythm rather than rapid bagging.

Not All BVM Devices Are Created Equal

The assumption in most training courses is that a BVM with a reservoir at high-flow oxygen will deliver somewhere near 90–100%. In practice, device design introduces enormous variability. A 2023 laboratory study tested 40 BVM devices from multiple manufacturers using a simulated spontaneously breathing patient primed with 100% oxygen at 15 liters per minute. Eight of the 40 devices couldn’t achieve an oxygen fraction above 0.85 for at least one tidal volume, and five models couldn’t hit that threshold at any tidal volume. Three devices delivered below 0.55, which is barely better than a simple face mask.³

The worst performers shared two design features: a duckbill-style non-rebreather valve and the absence of a dedicated expiratory valve. In those devices, oxygen concentration actually dropped as tidal volume increased, the opposite of what you’d want during vigorous breathing or ventilation. This means the device sitting in your crash cart or ambulance might not perform the way your textbook says it should, and without bench-testing the specific model, you wouldn’t know.

The Mask Seal Problem

Even a perfect BVM device is only as good as the seal between the mask and the patient’s face. A leak around the mask edges lets room air mix into the delivered breath and simultaneously lets oxygen-enriched gas escape into the room rather than entering the lungs. A preoxygenation study quantified this dramatically: a BVM held firmly on a mannequin face delivered a mean expired oxygen fraction of 76%, but with a simulated mask leak that number cratered to 30%, which is barely above room air.⁴ A 46-percentage-point drop just from an imperfect seal.

This is why two-person BVM technique is so strongly recommended. One person uses both hands to hold the mask against the face while the second person squeezes the bag. A study in preterm infants found that one-person mask ventilation produced a mean leak of about 26%, compared to roughly 18% with two-person technique.⁵ An earlier study of adult BVM ventilation with minimally trained rescuers was even more blunt, concluding that a single rescuer with a BVM failed to deliver adequate volumes for resuscitation, while two rescuers achieved tidal volumes comparable to endotracheal intubation.⁶

Facial anatomy adds another layer. Bearded patients, patients with no teeth (whose cheeks collapse inward), patients with facial trauma, and very small children all present seal challenges that no amount of squeezing force can fully compensate for. In these cases, oropharyngeal or nasopharyngeal airways inserted before mask ventilation help keep the airway open, but the seal issue remains a separate problem that technique alone has to solve.

Tidal Volume and the Risk of Giving Too Much

Oxygen concentration is only half the equation. The volume of each breath matters too. Adult BVM bags typically hold 1.5 liters or more, and a full squeeze delivers far more air than a patient needs. Current guidelines recommend tidal volumes of roughly 6–7 mL per kilogram of ideal body weight during resuscitation, which works out to about 400–600 mL for most adults. A study of emergency medical service providers found that with a standard adult BVM, the median delivered volume was around 835 mL, and fewer than 1% of breaths fell within the recommended range for a 70-kilogram patient. When providers switched to a smaller pediatric-sized bag for the same adult patient, about 46% of breaths landed in the target range.⁷

Hand size and technique compound the problem. Operators with larger hands delivered significantly higher volumes than those with smaller hands, and two-handed squeezes produced roughly 200–250 mL more per breath than one-handed squeezes across all hand sizes.⁸ That extra volume is desirable when a single operator is struggling to ventilate with one hand while maintaining a mask seal with the other. But in a two-person setup where the seal is already handled, a full two-handed squeeze of an adult bag can easily push volumes past 1,000 mL.⁹

Excessive volume creates real harm. High volumes generate high airway pressures, which can force air past the lower esophageal sphincter and into the stomach rather than the lungs. This gastric insufflation leads to vomiting and aspiration, which dramatically worsens outcomes. Bench models have shown that the sphincter opens at roughly 19 cm of water pressure, a threshold that vigorous bagging easily exceeds.¹⁰ The practical takeaway is to squeeze gently, use just enough volume to see the chest rise, and avoid the instinct to squeeze harder when ventilation feels difficult.

How BVM Compares to a Non-Rebreather Mask for Pre-Oxygenation

Before emergency intubation, clinicians “pre-oxygenate” the patient to build up an oxygen reserve in the lungs and blood, buying extra time during the intubation attempt. Both BVMs and non-rebreather masks (NRMs) are used for this purpose, and one question that comes up constantly is which one works better.

In a study comparing the two methods at flush-rate oxygen flow (meaning the flowmeter is opened all the way), the non-rebreather mask achieved a mean expired oxygen fraction of about 81%, while the BVM reached about 76%. The difference was small enough that the researchers judged the NRM to be non-inferior to the BVM for pre-oxygenation, meaning it worked about as well.⁴ The critical caveat, as noted earlier, is that the BVM’s performance collapsed when a mask leak was introduced. The NRM, by contrast, continued delivering in the low 80s even with an imperfect fit because its design doesn’t depend on an airtight seal the way a BVM does. For a conscious patient who is breathing on their own, an NRM may be the more reliable choice simply because it’s harder to screw up.

What Adding a PEEP Valve Does

A positive end-expiratory pressure (PEEP) valve attaches to the exhalation port of the BVM and keeps a small amount of pressure in the lungs at the end of each breath, preventing the small air sacs from collapsing completely. This matters for oxygen delivery because collapsed lung regions don’t participate in gas exchange, so keeping them open means the oxygen you deliver actually reaches the bloodstream more efficiently.

A bench and cadaver study found that adding a PEEP valve set at 5 cm of water pressure allowed eight out of nine BVM devices to deliver an expired oxygen fraction above 0.85.¹¹ A randomized trial in healthy volunteers compared BVM alone, BVM plus PEEP, and a non-rebreather mask for pre-oxygenation. In normal-weight adults, BVM plus PEEP achieved a mean expired oxygen fraction of about 76%, compared to 72% with BVM alone and 53% with the non-rebreather mask. The difference between BVM with and without PEEP was statistically significant. In overweight or obese adults, the gap was even wider: 73% with PEEP versus 66% without.¹² PEEP also improved ventilation in dependent (lower) lung regions, which are the first to collapse when someone is lying flat. For patients who are overweight or have stiff lungs, a PEEP valve may be the difference between adequate and inadequate pre-oxygenation.

Pediatric and Neonatal Setups Carry Extra Risks

Pediatric self-inflating resuscitators look like miniature versions of adult BVMs, but they have features that create unique failure modes. Most pediatric BVMs include a pop-off (pressure-relief) valve designed to prevent dangerously high airway pressures, and many also have a manometer port for measuring pressure during ventilation. The problem is that when these ports are left open, either by accident or by design flaw, delivered volumes drop substantially.

A bench study tested pediatric self-inflating resuscitators in both normal and poor lung compliance conditions. With proper setup (pop-off valve closed, manometer port capped), the devices delivered about 149 mL per breath under normal compliance. With the pop-off valve open, volume dropped to about 112 mL. With the manometer port open, it fell to about 106 mL. With both open, it cratered to 90 mL. Under poor compliance conditions, simulating a sick infant’s stiff lungs, the numbers were even more alarming: proper setup delivered about 122 mL, but with both ports open, only 44 mL reached the test lung.¹³ A volume that low may not ventilate a neonate effectively at all.

T-piece resuscitators, which are flow-inflating devices that require a compressed gas source, offer more precise pressure control for neonates. In a comparison of three neonatal resuscitation devices, the T-piece delivered target pressures more consistently than either the self-inflating bag or the flow-inflating bag. At a target peak pressure of 20 cm of water, the T-piece overshot by less than 1 cm on average, while the self-inflating bag overshot by nearly 5 cm.¹⁴ The trade-off is that T-pieces require a continuous gas supply and take longer to adjust pressures, which can be a disadvantage in rapidly changing clinical situations. The self-inflating bag also delivered significantly less PEEP than the T-piece, even when a PEEP valve was attached.

Provider Fatigue and Sustained Ventilation

BVM ventilation is physically tiring, especially when done single-handedly. A natural worry is that fatigue degrades the volumes and oxygen delivery over time as the rescuer’s grip weakens. One study measured tidal volumes during sustained manual ventilation across multiple time periods and found no significant decline due to fatigue alone. The effects of lung stiffness and whether one or two hands were used had a much larger impact on delivered volumes than how long the rescuer had been squeezing.¹⁵ That said, this was a controlled study. In a real resuscitation lasting 20 or 30 minutes, fatigue compounds with stress, distraction, and the physical awkwardness of holding a mask seal, all of which degrade performance in ways that are hard to capture in a lab.

When ventilating through an endotracheal tube rather than a mask, the seal problem disappears entirely, and both tidal volumes and airway pressures become more predictable. A comparison of BVM ventilation through an endotracheal tube versus a demand-valve device found that both produced acceptable tidal volumes and minute ventilation at normal and poor lung compliance, though the BVM generated higher peak airway pressures than the demand valve.¹⁶ Higher pressures increase the risk of barotrauma and gastric insufflation, so gentle, controlled squeezes remain important even when the airway is secured.

Aerosol Dispersion and Infection Risk

BVM ventilation generates aerosolized particles, a concern that became viscerally real during the COVID-19 pandemic. Every time gas leaks around the mask seal, it carries respiratory droplets and aerosols outward toward the healthcare provider. A study using human subjects and smoke visualization found that exhaled air leaked an average of about 267 mm (roughly 10 inches) from the mask when nurses used a silicone resuscitator. Interestingly, nurses produced larger leak distances than anesthesiologists, respiratory physicians, or medical students, who all performed similarly to each other. The difference likely reflects hand size and grip technique rather than clinical knowledge.¹⁷

Newer ventilation strategies aim to reduce aerosol exposure. A comparison of standard bag-valve ventilation against chest compression-synchronized ventilation (CCSV) during simulated CPR found that after five minutes, aerosol particle counts were significantly lower with the synchronized approach. With standard BVM technique, aerosol counts continued climbing throughout the experiment, while CCSV reached a plateau and stayed there.¹⁸ For providers working without powered air-purifying respirators, this kind of aerosol reduction could meaningfully lower the risk of acquiring a respiratory infection during resuscitation. Inline viral/bacterial filters placed between the BVM and the mask or endotracheal tube add another layer of protection, though they introduce a small amount of dead space that can slightly reduce the effective tidal volume reaching the patient.