You're rounding on a COPD patient at 3 AM. They're on 4 liters via nasal cannula. The nurse asks, "So what's their FiO2 right now?
You hesitate. Because the honest answer is: nobody actually knows Still holds up..
What Is Low-Flow Oxygen Delivery
Low-flow oxygen means the device delivers less gas flow than the patient's peak inspiratory demand. That's the technical definition. In practice, it means nasal cannulas, simple face masks, and those weird little pediatric hoods nobody uses anymore.
Here's the thing — the oxygen source (wall outlet, tank, concentrator) puts out 100% O2. But the patient doesn't breathe 100% O2. So naturally, they breathe a mixture. On the flip side, room air gets entrained around the edges of the cannula prongs, under the mask seal, through the vents. The final inspired concentration — the FiO2 — depends entirely on how much room air gets pulled in alongside that pure oxygen.
And that varies. Wildly And that's really what it comes down to..
The variable performance problem
Low-flow devices are technically called "variable performance" devices. And fixed performance (Venturi masks, high-flow nasal cannula) give you a known, reliable FiO2. Low-flow? In real terms, you're guessing. But educated guessing, maybe. But guessing.
The classic teaching: nasal cannula at 1 L/min ≈ 24% FiO2. 2 L/min ≈ 28%. Here's the thing — 3 L/min ≈ 32%. 4 L/min ≈ 36%. Which means 5 L/min ≈ 40%. 6 L/min ≈ 44%.
Neat. Memorable. Wrong for any given patient at any given moment.
Why It Matters
You might think, "Close enough. If they're saturating 94%, who cares about the exact number?"
Until it isn't close enough And that's really what it comes down to..
The CO2 retainer trap
That COPD patient at 3 AM? " Their sats come up to 98%. They're a CO2 retainer. Consider this: you crank the cannula to 6 L/min because "they need more oxygen. Plus, their respiratory drive is hypoxic, not hypercapnic. You feel good.
Two hours later they're obtunded. pH 7.18. PaCO2 92.
You delivered way more FiO2 than you realized — maybe 50%, maybe 60% — because their tidal volume dropped, their respiratory rate slowed, and suddenly that 6 L/min was meeting a huge chunk of their inspiratory demand. Less room air entrainment. Higher FiO2. Blunted hypoxic drive. Disaster Nothing fancy..
The "they look fine" trap
Flip side: young, healthy post-op patient. 2 L/min nasal cannula. That 2 L/min is a drop in the bucket. Their actual FiO2? You assume FiO2 ~28%. Barely 24%. Now, sats 97%. They're working hard to breathe room air with a tiny oxygen supplement. You miss the atelectasis. But they're anxious, tachypneic, pulling 40 L/min minute ventilation. In practice, you miss the early shunt. You send them home and they bounce back in 12 hours.
The number matters. Not for the chart. For the clinical decision.
How It Actually Works
Let's talk physics. Don't worry — no equations you can't do in your head Easy to understand, harder to ignore..
The mixing chamber is the nasopharynx
Nasal cannula prongs sit in the nares. Oxygen flows out continuously. During exhalation, the nasopharynx fills with 100% O2 — a reservoir. During inspiration, the patient pulls gas from that reservoir plus room air entrained through the mouth and around the prongs Not complicated — just consistent..
The ratio determines FiO2.
FiO2 = (O2 flow + entrained air × 0.21) / total inspired flow
Since entrained air is 21% oxygen, the math simplifies to: every liter of oxygen flow adds roughly 3-4% FiO2 above room air — but only if the patient's inspiratory flow is low enough that the reservoir doesn't get depleted It's one of those things that adds up. Turns out it matters..
Minute ventilation is the variable you can't control
A normal adult at rest: tidal volume ~500 mL, rate ~12, minute ventilation ~6 L/min. Peak inspiratory flow? Maybe 30 L/min.
Your 4 L/min cannula meets 13% of peak demand. And the other 87% is room air. FiO2 ends up around 32-36% The details matter here..
Same patient, pneumonia, anxious, tachypneic: rate 28, tidal volume 600 mL, minute ventilation ~17 L/min. Think about it: peak inspiratory flow 60+ L/min. Now your 4 L/min meets 6-7% of peak demand. FiO2 drops to ~26% Small thing, real impact..
Same patient, oversedated: rate 6, tidal volume 300 mL, minute ventilation ~1.8 L/min. Even so, peak flow maybe 15 L/min. Worth adding: your 4 L/min meets 25%+ of demand. FiO2 climbs toward 45-50% It's one of those things that adds up..
The oxygen flow didn't change. The patient did That's the part that actually makes a difference..
Mouth breathing changes everything
Nasal cannula assumes nasal breathing. Mouth open? In practice, the reservoir effect vanishes. Room air pours in unobstructed. FiO2 drops like a stone — often 5-10% lower than predicted.
Simple face masks? Those side vents are huge. Consider this: same problem. At 5-6 L/min you might get 40-50% FiO2 if the seal is good and breathing is quiet. Cough, talk, turn over — the seal breaks, FiO2 plummets.
The reservoir effect has limits
That nasopharyngeal reservoir holds maybe 50-100 mL in an adult. At high inspiratory flows, it empties in milliseconds. The rest of the breath is pure room air.
This is why the "1 L = 4%" rule breaks down above 6 L/min. Noise. Still, waste. Also, you can't fit more reservoir. You just blow oxygen past the patient's face. In practice, dry mucosa. No added benefit.
Common Mistakes
Treating the chart like a contract
"Patient on 4 L NC = 36% FiO2.That said, " Written in the orders. Signed off in the handoff. Treated as truth Simple, but easy to overlook..
It's not truth. Day to day, it's a rough estimate for a typical adult at rest. Your patient is rarely typical and rarely at rest Most people skip this — try not to..
Assuming more flow = linear increase
It's not linear. Now, the curve flattens hard above 4-5 L/min on nasal cannula. On the flip side, going from 5 to 6 L/min might buy you 2% FiO2. And going from 2 to 3 L/min buys 4%. Diminishing returns are real.
Ignoring the humidification factor
Dry gas at 6 L/min destroys nasal mucosa in hours. Crusting, bleeding
and irritation. This leads to more coughing, which increases work of breathing, which increases inspiratory flow, which—as we have established—lowers the FiO2. It is a physiological feedback loop that works against your goal.
Clinical Implications: The "Oxygen Trap"
Because of these variables, clinicians often fall into the trap of increasing flow to solve a desaturation, without addressing the underlying cause of the increased demand. If a patient’s SpO2 drops from 94% to 88%, the instinct is to turn the dial from 2 L to 4 L Simple, but easy to overlook. That alone is useful..
Still, if that drop is due to the patient becoming tachypneic (increasing their peak inspiratory flow), the jump to 4 L might only increase their FiO2 by a fraction of a percent. You are essentially throwing more gas into a furnace that is already wide open, without actually increasing the temperature That's the part that actually makes a difference. Simple as that..
Summary: Moving from Math to Physiology
To provide effective oxygen therapy, you must stop viewing the flow meter as a precision instrument and start viewing it as a tool for managing a highly volatile biological system.
- Understand the "Why": If a patient's FiO2 is dropping despite a high flow rate, look at their breathing pattern. Are they tachypneic? Are they mouth-breathing? Are they fighting the equipment?
- Monitor the Patient, Not the Dial: The pulse oximetry reading is a reflection of the patient's work of breathing and gas exchange, not just the number on the flow meter.
- Know the Limits: Recognize that nasal cannulas have a ceiling. Once you hit the point of diminishing returns, it is time to escalate to a higher-delivery device (like a Venturi mask or High-Flow Nasal Cannula) rather than simply cranking up the flow on a simple cannula.
All in all, oxygen therapy is a balancing act between the physics of gas flow and the unpredictable mechanics of human respiration. Mastery of oxygenation requires moving beyond the "rule of thumb" math and developing a clinical intuition for how a patient's breathing pattern dictates the actual efficacy of the therapy provided.
